feat: initialize managed portal

This commit is contained in:
Yoilun
2026-04-27 10:04:36 +08:00
commit d4e351df71
145 changed files with 13425 additions and 0 deletions

View File

@@ -0,0 +1,7 @@
.git
.venv
__pycache__
*.pyc
outputs
people_flow_project_backup_2026-04-08
docs/plans

View File

@@ -0,0 +1,8 @@
.DS_Store
.venv/
__pycache__/
config/local.yaml
outputs/
wheelhouse/
weights/*.pt
weights/deepface/*.h5

View File

@@ -0,0 +1,44 @@
FROM swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/library/python:3.12-slim-bookworm
ENV PYTHONDONTWRITEBYTECODE=1 \
PYTHONUNBUFFERED=1 \
PIP_NO_CACHE_DIR=1 \
PIP_INDEX_URL=https://pypi.tuna.tsinghua.edu.cn/simple \
DEEPFACE_HOME=/root/.deepface \
TF_CPP_MIN_LOG_LEVEL=2
WORKDIR /opt/people-flow
RUN apt-get update && apt-get install -y --no-install-recommends \
ca-certificates \
libglib2.0-0 \
libgl1 \
libgomp1 \
&& rm -rf /var/lib/apt/lists/*
COPY requirements-docker.txt ./requirements-docker.txt
RUN python -m pip install --upgrade pip setuptools wheel && \
pip install "numpy<2" && \
pip install --extra-index-url https://download.pytorch.org/whl/cpu \
"torch==2.6.0+cpu" "torchvision==0.21.0+cpu" && \
pip install "tensorflow==2.16.1" "tf-keras==2.16.0" && \
pip install -r requirements-docker.txt
COPY . .
COPY scripts/docker-entrypoint.sh /opt/people-flow/scripts/docker-entrypoint.sh
RUN test -f /opt/people-flow/weights/yolo11n.pt && \
test -f /opt/people-flow/weights/deepface/age_model_weights.h5 && \
test -f /opt/people-flow/weights/deepface/gender_model_weights.h5 && \
test -f /opt/people-flow/weights/deepface/retinaface.h5 && \
mkdir -p /root/.deepface/weights /opt/people-flow/outputs && \
cp /opt/people-flow/weights/deepface/*.h5 /root/.deepface/weights/ && \
chmod +x /opt/people-flow/scripts/docker-entrypoint.sh
EXPOSE 18082
HEALTHCHECK --interval=30s --timeout=5s --start-period=15s --retries=3 \
CMD python -c "import urllib.request; urllib.request.urlopen('http://127.0.0.1:18082/api/manage/health', timeout=3).read()" || exit 1
ENTRYPOINT ["/opt/people-flow/scripts/docker-entrypoint.sh"]

View File

@@ -0,0 +1,144 @@
# People Flow Project
People flow analysis for street videos using YOLO tracking and face-based demographic estimation.
## What it does
- Counts unique people when they cross a configured line
- Estimates one age bucket per counted track: `minor`, `adult`, or `senior`
- Estimates one gender bucket per counted track: `male` or `female`
- Writes an annotated output video, per-video JSON, and batch summary CSV
## Pipeline
1. Detect and track `person` objects with Ultralytics YOLO.
2. Assign a stable `track_id` with BoT-SORT or ByteTrack.
3. Count each track once when it crosses the configured line.
4. Sample person crops for each track and run DeepFace age/gender analysis.
5. Use track-level voting so each counted person lands in only one age bucket and one gender bucket.
## Project Layout
- `main.py`: CLI entrypoint
- `src/people_flow/`: application modules
- `configs/default_config.yaml`: default runtime settings
- `outputs/`: generated result files
- `docs/plans/`: design and implementation notes
## Recommended Environment
- Linux
- NVIDIA GPU with CUDA
- Python `3.10` or `3.11`
`deepface` and its transitive dependencies are not a good fit for Python `3.14`, so do not build this environment on the current local interpreter version.
## Install
```bash
python3.11 -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt
```
## Single Video
```bash
python main.py video \
--input "/path/to/video.mp4" \
--line "0.1,0.55,0.9,0.55"
```
## Batch Directory
```bash
python main.py batch \
--input-dir "/path/to/videos" \
--line "0.1,0.55,0.9,0.55"
```
## RTSP Stream
```bash
python main.py --output-dir outputs rtsp \
--input "rtsp://user:password@host:554/stream"
```
RTSP mode behaves differently from offline video mode:
- The stream is sampled at one processed frame per second
- Statistics are isolated into 30-minute windows
- Each completed window writes one JSON file
- `latest.json` is overwritten on every completed window
- RTSP mode does not save annotated video by default
## Output Files
Each processed video produces:
- `outputs/<video_stem>/<video_stem>.annotated.mp4`
- `outputs/<video_stem>/<video_stem>.json`
Batch mode also produces:
- `outputs/batch_summary.csv`
RTSP mode produces:
- `outputs/rtsp_stream/latest.json`
- `outputs/rtsp_stream/windows/stats_YYYY-MM-DD_HH-MM-SS.json`
## Docker On Ubuntu 24.04 x86_64
The project can be packaged for an x86_64 NVIDIA host with Docker. The expected weight layout is:
- `weights/yolo11n.pt`
- `weights/deepface/age_model_weights.h5`
- `weights/deepface/gender_model_weights.h5`
- `weights/deepface/retinaface.h5`
Build the image:
```bash
docker build -t people-flow-project:test .
```
The Docker image uses [`requirements-docker.txt`](/Users/zxmacmini1/Documents/人流检测/people_flow_project/requirements-docker.txt) so the container installs `opencv-python-headless` instead of the desktop OpenCV wheel.
The image bakes in all runtime weights and copies the DeepFace `.h5` files into `~/.deepface/weights` during build.
Run the management API container:
```bash
docker run -d \
--name people-flow-project \
--restart unless-stopped \
--gpus all \
--shm-size 1g \
-p 18082:18082 \
-e RTSP_URL="rtsp://user:password@host:554/stream" \
-v /path/to/config:/opt/people-flow/config \
-v /path/to/outputs:/opt/people-flow/outputs \
people-flow-project:test
```
Or use Compose:
```bash
docker compose up --build people-flow-project
```
Container behavior:
- Seeds `config/local.yaml` from `config/config.example.yaml` when needed
- Writes RTSP updates through the child API to `runtime.rtsp_url`
- Exposes `GET /api/manage/health` on `http://127.0.0.1:18082`
- Persists config and outputs through mounted `./config` and `./outputs`
## Notes
- `minor` means age `< 18`
- `adult` means age `18-59`
- `senior` means age `>= 60`
- Tracks without a reliable face result are counted only in `total_people` and `unknown_attributes`

View File

@@ -0,0 +1,65 @@
# Native RTSP Bundle
This bundle is for lightweight native deployment on an x86_64 Ubuntu host.
## What To Edit
Open [`scripts/run.sh`](/Users/zxmacmini1/Documents/人流检测/people_flow_project/scripts/run.sh) and edit only these two lines:
```bash
RTSP_URL="rtsp://..."
OUTPUT_DIR="/home/x/people/output"
```
## First-Time Setup
From the project root:
```bash
sudo bash scripts/install.sh
```
This creates `.venv`, installs Python dependencies, copies the bundled DeepFace weights into `~/.deepface/weights`, installs the `systemd` unit, starts the service, and enables it on boot.
## Build An Offline Dependency Pack
If you want future installs to avoid re-downloading Python packages:
```bash
./build_wheelhouse.sh
```
This creates a local `wheelhouse/` directory for Ubuntu 24.04 x86_64 + Python 3.12. After that, `./setup_native_venv.sh` will automatically prefer local wheels.
## Start The RTSP Task
```bash
sudo systemctl status people-flow.service
```
The service runs in the foreground under `systemd`.
## Outputs
- Latest half-hour summary: `OUTPUT_DIR/rtsp_stream/latest.json`
- Historical half-hour summaries: `OUTPUT_DIR/rtsp_stream/windows/`
- Runtime log: `OUTPUT_DIR/rtsp_run.log`
## Chinese Guide
- `README_zh.md`
## Weights
The project expects these local files:
- `weights/yolo11n.pt`
- `weights/deepface/age_model_weights.h5`
- `weights/deepface/gender_model_weights.h5`
- `weights/deepface/retinaface.h5`
At setup time and each RTSP launch, those `.h5` files are copied into the current user's default DeepFace directory:
- `~/.deepface/weights/`
That keeps the bundle portable across different unpack paths such as `/home/x/people` and `/home/xiaozheng/people`.

View File

@@ -0,0 +1,72 @@
# 人流检测项目中文说明
这个项目用于基于 `YOLO + DeepFace` 的视频/RTSP 人流检测与属性统计。
## 当前交付方式
这个版本已经改成:
- 使用 `config/local.yaml` 作为本地运行配置
- 使用 `scripts/run.sh` 生成本地配置并前台运行
- 使用 `systemd` 托管长期运行
- 安装完成后自动启动
- 开机自动启动
## 目标机器
- `Ubuntu 24.04`
- `Python 3.12`
- NVIDIA 显卡可用
- `nvidia-smi` 可正常执行
## 安装前需要修改
先编辑 `scripts/run.sh`,至少改:
- `RTSP_URL`
- `OUTPUT_DIR`
## 安装
在项目根目录执行:
```bash
sudo bash scripts/install.sh
```
安装脚本会自动:
- 检查并安装 `ffmpeg`
- 检查并安装 `python3.12-venv`
- 创建 `.venv`
- 安装 Python 依赖
- 复制 DeepFace 权重到 `~/.deepface/weights`
- 生成 `config/local.yaml`
- 安装 `systemd` 服务
- 自动启动服务
- 设置开机自启
## 服务管理
服务名:
```bash
people-flow.service
```
常用命令:
```bash
sudo systemctl status people-flow.service
sudo systemctl restart people-flow.service
sudo systemctl stop people-flow.service
sudo systemctl start people-flow.service
sudo systemctl disable people-flow.service
```
## 输出位置
- 运行日志:`outputs/rtsp_run.log`
- 最新半小时汇总:`OUTPUT_DIR/rtsp_stream/latest.json`
- 历史窗口汇总:`OUTPUT_DIR/rtsp_stream/windows/`
- 本地配置:`config/local.yaml`

View File

@@ -0,0 +1,31 @@
#!/usr/bin/env bash
set -euo pipefail
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
PROJECT_ROOT="$SCRIPT_DIR"
WHEELHOUSE_DIR="$PROJECT_ROOT/wheelhouse"
mkdir -p "$WHEELHOUSE_DIR"
python3 -m venv "$PROJECT_ROOT/.wheelhouse-venv"
source "$PROJECT_ROOT/.wheelhouse-venv/bin/activate"
python -m pip install --upgrade pip setuptools wheel
pip download -d "$WHEELHOUSE_DIR" pip setuptools wheel
pip download -d "$WHEELHOUSE_DIR" "numpy<2"
pip download -d "$WHEELHOUSE_DIR" \
--index-url https://download.pytorch.org/whl/cu126 \
--extra-index-url https://pypi.nvidia.com \
torch torchvision
pip download -d "$WHEELHOUSE_DIR" \
--extra-index-url https://pypi.nvidia.com \
"tensorflow[and-cuda]==2.16.1" "tf-keras==2.16.0"
pip download -d "$WHEELHOUSE_DIR" \
--find-links "$WHEELHOUSE_DIR" \
-c "$PROJECT_ROOT/constraints-wheelhouse.txt" \
-r "$PROJECT_ROOT/requirements-native.txt"
deactivate
echo "wheelhouse_ready=$WHEELHOUSE_DIR"

View File

@@ -0,0 +1,41 @@
runtime:
rtsp_url: "rtsp://user:password@camera-ip:554/h264/ch1/main/av_stream"
output_dir: "outputs"
yolo:
model_path: "weights/yolo11n.pt"
tracker: "botsort.yaml"
conf: 0.35
iou: 0.5
imgsz: 1280
device: "cuda:0"
counting:
line: [0.1, 0.55, 0.9, 0.55]
line_mode: "normalized"
crossing_tolerance: 12.0
attributes:
enabled: false
sample_every_n_frames: 12
max_samples_per_track: 5
min_person_box_width: 80
min_person_box_height: 160
person_crop_padding: 0.15
detector_backend: "retinaface"
enforce_detection: false
output:
save_video: false
save_json: true
save_csv: true
draw_boxes: false
draw_labels: false
rtsp:
sample_interval_seconds: 1.0
window_seconds: 1800
reconnect_delay_seconds: 5.0
stream_open_timeout_seconds: 10.0
idle_sleep_seconds: 0.05
output_subdir: "rtsp_stream"

View File

@@ -0,0 +1,37 @@
yolo:
model_path: "yolo11n.pt"
tracker: "botsort.yaml"
conf: 0.35
iou: 0.5
imgsz: 1280
device: "cuda:0"
counting:
line: [0.1, 0.55, 0.9, 0.55]
line_mode: "normalized"
crossing_tolerance: 12.0
attributes:
enabled: true
sample_every_n_frames: 12
max_samples_per_track: 5
min_person_box_width: 80
min_person_box_height: 160
person_crop_padding: 0.15
detector_backend: "retinaface"
enforce_detection: false
output:
save_video: true
save_json: true
save_csv: true
draw_boxes: true
draw_labels: true
rtsp:
sample_interval_seconds: 1.0
window_seconds: 1800
reconnect_delay_seconds: 5.0
stream_open_timeout_seconds: 10.0
idle_sleep_seconds: 0.05
output_subdir: "rtsp_stream"

View File

@@ -0,0 +1,37 @@
yolo:
model_path: "/opt/people-flow/weights/yolo11n.pt"
tracker: "botsort.yaml"
conf: 0.35
iou: 0.5
imgsz: 1280
device: "cuda:0"
counting:
line: [0.1, 0.55, 0.9, 0.55]
line_mode: "normalized"
crossing_tolerance: 12.0
attributes:
enabled: true
sample_every_n_frames: 12
max_samples_per_track: 5
min_person_box_width: 80
min_person_box_height: 160
person_crop_padding: 0.15
detector_backend: "retinaface"
enforce_detection: false
output:
save_video: false
save_json: true
save_csv: true
draw_boxes: false
draw_labels: false
rtsp:
sample_interval_seconds: 1.0
window_seconds: 1800
reconnect_delay_seconds: 5.0
stream_open_timeout_seconds: 10.0
idle_sleep_seconds: 0.05
output_subdir: "rtsp_stream"

View File

@@ -0,0 +1,37 @@
yolo:
model_path: "weights/yolo11n.pt"
tracker: "botsort.yaml"
conf: 0.35
iou: 0.5
imgsz: 1280
device: "cuda:0"
counting:
line: [0.1, 0.55, 0.9, 0.55]
line_mode: "normalized"
crossing_tolerance: 12.0
attributes:
enabled: true
sample_every_n_frames: 12
max_samples_per_track: 5
min_person_box_width: 80
min_person_box_height: 160
person_crop_padding: 0.15
detector_backend: "retinaface"
enforce_detection: false
output:
save_video: false
save_json: true
save_csv: true
draw_boxes: false
draw_labels: false
rtsp:
sample_interval_seconds: 1.0
window_seconds: 1800
reconnect_delay_seconds: 5.0
stream_open_timeout_seconds: 10.0
idle_sleep_seconds: 0.05
output_subdir: "rtsp_stream"

View File

@@ -0,0 +1,5 @@
numpy<2
tensorflow==2.16.1
tf-keras==2.16.0
torch==2.11.0+cu126
torchvision==0.26.0+cu126

View File

@@ -0,0 +1,19 @@
[Unit]
Description=People Flow RTSP Service
After=network-online.target
Wants=network-online.target
[Service]
Type=simple
WorkingDirectory=__PROJECT_DIR__
User=__RUN_USER__
Group=__RUN_GROUP__
Environment=PYTHONUNBUFFERED=1
ExecStart=__PROJECT_DIR__/.venv/bin/python __PROJECT_DIR__/main.py --config __CONFIG_PATH__ rtsp
Restart=always
RestartSec=5
StandardOutput=append:__PROJECT_DIR__/outputs/rtsp_run.log
StandardError=append:__PROJECT_DIR__/outputs/rtsp_run.log
[Install]
WantedBy=multi-user.target

View File

@@ -0,0 +1,22 @@
services:
people-flow-project:
build:
context: .
dockerfile: Dockerfile
image: people-flow-project:local
container_name: people-flow-project
restart: unless-stopped
gpus: all
shm_size: "1gb"
ports:
- "18082:18082"
environment:
CONFIG_PATH: /opt/people-flow/config/local.yaml
RTSP_URL: ${RTSP_URL:-}
OUTPUT_DIR: /opt/people-flow/outputs
API_HOST: 0.0.0.0
API_PORT: 18082
DEVICE: ${DEVICE:-cuda:0}
volumes:
- ./config:/opt/people-flow/config
- ./outputs:/opt/people-flow/outputs

View File

@@ -0,0 +1,66 @@
# People Flow Design
## Goal
Build a standalone project under `Documents/人流检测/people_flow_project` that analyzes street videos and produces:
- unique people-flow counts
- one mutually exclusive age bucket per counted person
- one mutually exclusive gender bucket per counted person
- annotated videos plus machine-readable summaries
## Approved Decisions
- Runtime target: Linux with NVIDIA GPU
- Entry points: both single-video mode and batch-directory mode
- Count logic: one `track_id` is counted once when it crosses a configured line
- Age buckets:
- `minor`: age `< 18`
- `adult`: age `18-59`
- `senior`: age `>= 60`
- Gender buckets:
- `male`
- `female`
- Unknown face attributes:
- If a counted person does not yield a reliable face result, count that person only in `total_people`
- Also increment `unknown_attributes`
## Architecture
The pipeline uses Ultralytics YOLO for person detection and tracking, then DeepFace for face attribute analysis. Person tracking and counting stay separate from attribute inference so the demographic model can be replaced later without touching the counting core.
The application stores votes per `track_id`. When the video finishes, each counted track is resolved to at most one final age bucket and one final gender bucket by majority voting.
## Modules
- `main.py`: CLI parsing and mode dispatch
- `src/people_flow/config.py`: config loading and overrides
- `src/people_flow/tracking.py`: track extraction from YOLO results
- `src/people_flow/counting.py`: line-crossing logic and unique counting
- `src/people_flow/attributes.py`: DeepFace integration and voting
- `src/people_flow/io_utils.py`: video, JSON, and CSV output helpers
- `src/people_flow/pipeline.py`: process orchestration
## Outputs
For each video:
- annotated MP4
- JSON summary
For batch runs:
- one CSV summary with one row per video
## Error Handling
- Missing dependencies should raise clear installation guidance.
- If a video cannot be opened, fail that video with a readable error.
- If face inference fails for a sample, continue processing and treat that sample as unavailable.
- If no video files are found in batch mode, fail fast with a clear message.
## Limitations
- Age and gender quality depend on clear, sufficiently large faces.
- Street scenes with strong occlusion, side views, masks, or low light will increase `unknown_attributes` and lower reliability.
- The default line is a placeholder and should be adjusted per camera view.

View File

@@ -0,0 +1,128 @@
# People Flow Implementation Plan
> **For Claude:** REQUIRED SUB-SKILL: Use superpowers:executing-plans to implement this plan task-by-task.
**Goal:** Build a standalone Python project that counts unique line crossings in street videos and adds track-level age/gender summaries.
**Architecture:** Use Ultralytics YOLO to detect and track persons frame by frame, then run DeepFace on sampled person crops to infer face attributes. Keep counting, tracking, and attribute voting in separate modules so the demographic backend can be swapped later.
**Tech Stack:** Python, Ultralytics YOLO, OpenCV, DeepFace, PyYAML, pandas
---
### Task 1: Scaffold the project
**Files:**
- Create: `README.md`
- Create: `requirements.txt`
- Create: `pyproject.toml`
- Create: `configs/default_config.yaml`
- Create: `docs/plans/2026-04-07-people-flow-design.md`
**Step 1: Write the initial files**
Add installation instructions, runtime expectations, and default settings.
**Step 2: Verify structure**
Run: `find . -maxdepth 3 | sed -n '1,120p'`
Expected: project files and directories exist.
**Step 3: Commit**
This workspace is not a git repository. Skip the commit step unless the user later initializes git here.
### Task 2: Build the CLI and config loader
**Files:**
- Create: `main.py`
- Create: `src/people_flow/__init__.py`
- Create: `src/people_flow/config.py`
- Create: `src/people_flow/models.py`
**Step 1: Implement argument parsing**
Support `video` and `batch` subcommands, config overrides, output directory selection, and line overrides.
**Step 2: Implement config loading**
Load YAML defaults and merge CLI overrides into typed dataclasses.
**Step 3: Verify**
Run: `python3 -m compileall main.py src`
Expected: compile succeeds without syntax errors.
### Task 3: Implement tracking and counting
**Files:**
- Create: `src/people_flow/tracking.py`
- Create: `src/people_flow/counting.py`
**Step 1: Extract tracked `person` detections**
Convert YOLO result objects into simple track observations with `track_id`, bounding box, confidence, and center point.
**Step 2: Implement line-cross counting**
Count one crossing per track by monitoring the sign change of the track center relative to the configured line.
**Step 3: Verify**
Run: `python3 -m compileall src`
Expected: compile succeeds.
### Task 4: Implement attribute voting and output helpers
**Files:**
- Create: `src/people_flow/attributes.py`
- Create: `src/people_flow/io_utils.py`
**Step 1: Integrate DeepFace**
Sample person crops, run `age` and `gender` analysis, normalize labels, and store per-track votes.
**Step 2: Implement output helpers**
Write JSON summaries, CSV summaries, and draw overlays onto frames.
**Step 3: Verify**
Run: `python3 -m compileall src`
Expected: compile succeeds.
### Task 5: Implement the processing pipeline
**Files:**
- Create: `src/people_flow/pipeline.py`
**Step 1: Build the main loop**
Open the video, run YOLO tracking on frames, update counters, sample attributes, draw overlays, and save artifacts.
**Step 2: Build batch mode**
Discover supported video files recursively and run the same pipeline per file, then write `outputs/batch_summary.csv`.
**Step 3: Verify**
Run: `python3 -m compileall main.py src`
Expected: compile succeeds.
### Task 6: Final verification
**Files:**
- Modify: `README.md`
**Step 1: Smoke-check the CLI**
Run: `python3 main.py --help`
Expected: help text shows the `video` and `batch` commands.
**Step 2: Document limitations**
Make sure README notes Python version constraints and face-quality limitations.
**Step 3: Commit**
Skip commit because this workspace is not a git repository.

View File

@@ -0,0 +1,17 @@
# Portable DeepFace Weights Design
**Goal:** Make DeepFace reuse bundled project weights regardless of where the project directory is unpacked.
**Problem:** The current native launcher sets `DEEPFACE_HOME` to a project-local `.deepface` directory. DeepFace then appends its own `.deepface/weights` segment, so runtime lookup becomes `PROJECT_ROOT/.deepface/.deepface/weights`, which bypasses the bundled `weights/deepface` directory and triggers redundant downloads.
**Approach Options:**
1. Copy bundled weights into the current user's default `~/.deepface/weights` directory before startup.
This matches DeepFace's default lookup behavior and avoids hard-coded absolute paths. It works whether the project lives under `/home/x/people`, `/home/xiaozheng/people`, or any other directory.
2. Keep using `DEEPFACE_HOME` and reshape the project-local directory tree to match DeepFace's nested expectations.
This avoids duplicating files but is more fragile and easier to break when DeepFace internals change.
**Recommendation:** Use option 1. Update the native setup and launcher scripts to sync `weights/deepface/*.h5` into `~/.deepface/weights` and stop overriding `DEEPFACE_HOME`.
**Validation:** Confirm the RTSP process starts without downloading `retinaface.h5`, `age_model_weights.h5`, or `gender_model_weights.h5`, and verify the launcher still works after changing only the project root path.

View File

@@ -0,0 +1,51 @@
# Portable DeepFace Weights Implementation Plan
> **For Claude:** REQUIRED SUB-SKILL: Use superpowers:executing-plans to implement this plan task-by-task.
**Goal:** Make the native RTSP bundle reuse bundled DeepFace weights from any unpack location without extra downloads.
**Architecture:** Remove the custom `DEEPFACE_HOME` override from the native runtime path. Before setup and launch, copy the bundled DeepFace weight files from `weights/deepface/` into the current user's default `~/.deepface/weights/` directory so DeepFace resolves them through its own standard path logic.
**Tech Stack:** Bash, DeepFace, native Python virtual environment, offline wheelhouse bundle
---
### Task 1: Fix native setup and launcher paths
**Files:**
- Modify: `run_rtsp.sh`
- Modify: `setup_native_venv.sh`
- Modify: `README_NATIVE.md`
**Step 1: Update `run_rtsp.sh`**
Remove the `DEEPFACE_HOME` override. Create `"$HOME/.deepface/weights"` and copy bundled `.h5` files from `"$PROJECT_ROOT/weights/deepface"` into that directory before starting the Python process.
**Step 2: Update `setup_native_venv.sh`**
After dependency installation, create `"$HOME/.deepface/weights"` and copy bundled `.h5` files into it so the environment is ready before the first run.
**Step 3: Update native documentation**
Explain that bundled weights are staged into `~/.deepface/weights` automatically and that the project path itself can move without breaking the weight lookup.
### Task 2: Sync and verify on the Ubuntu target
**Files:**
- Modify: remote copies of the files above under `/home/x/people/people_flow_project`
**Step 1: Sync the changed files to `192.168.5.154`**
Copy the updated launcher, setup script, and documentation.
**Step 2: Stage bundled weights into the target user's home directory**
Run the updated setup logic or equivalent copy command and verify `~/.deepface/weights` contains the expected `.h5` files.
**Step 3: Restart RTSP and inspect logs**
Restart the RTSP job and confirm the log no longer shows downloads from `deepface_models/releases`.
**Step 4: Commit**
Skip commit unless explicitly requested by the user.

View File

@@ -0,0 +1,81 @@
# Lightweight Native Bundle Design
**Date:** 2026-04-08
**Goal:** Deliver a lightweight native deployment bundle for Ubuntu 24.04 x86_64 that includes project code, required weights, a single editable RTSP run script, and a small setup path on the target host without bundling a full Python environment in the archive.
## Scope
- Target host: `xiaozheng@192.168.5.154`
- Target path: `/home/x/people/people_flow_project`
- Bundle contents:
- project code
- YOLO weight
- DeepFace weights
- one editable run script
- setup and usage documentation
- Exclude the virtual environment from the compressed bundle to keep size down.
## Deployment Model
The target host already has:
- Ubuntu 24.04 x86_64
- Python 3.12
- Docker available, but Docker is intentionally not used here
- NVIDIA driver and CUDA-capable GPU
The bundle will therefore rely on:
1. a project-local `.venv` created on the target host
2. host driver compatibility for GPU wheels
3. project-relative weight paths so no external downloads are needed
## User Editing Surface
The main operator interface is a single shell script:
- `run_rtsp.sh`
The user edits only:
- `RTSP_URL`
- `OUTPUT_DIR`
The script activates `.venv`, points to the native x86 config, and runs the RTSP pipeline.
## Config Strategy
Add a dedicated native x86 config file with:
- `yolo.model_path` pointing to the local `weights/yolo11n.pt`
- RTSP timing settings
- output defaults for RTSP mode
This avoids modifying the existing Jetson-oriented config and keeps host deployment deterministic.
## Setup Strategy
Provide a small setup script that:
- creates `.venv`
- upgrades pip/setuptools/wheel
- installs CUDA-enabled PyTorch wheels
- installs TensorFlow, `tf-keras`, and application dependencies
The setup script keeps the archive light while still making the target directory self-contained after one install step.
## Bundle Output
On the target host, create a compressed archive such as:
- `/home/x/people/people_flow_project_native_bundle_2026-04-08.tar.gz`
The archive will exclude `.venv` so it stays close to the size of code plus weights.
## Success Criteria
- The target host contains a runnable native project directory
- `run_rtsp.sh` is the only file the operator needs to edit for RTSP URL and output directory
- All required weights are present locally
- The lightweight tarball is created successfully

View File

@@ -0,0 +1,66 @@
# Lightweight Native Bundle Implementation Plan
> **For Claude:** REQUIRED SUB-SKILL: Use superpowers:executing-plans to implement this plan task-by-task.
**Goal:** Produce a lightweight native deployment bundle for Ubuntu 24.04 x86_64 with code, weights, one editable RTSP run script, and a local venv setup path.
**Architecture:** Keep all code and weights inside the project directory, add one native config and two helper scripts, then create the venv on the target host instead of bundling it into the archive.
**Tech Stack:** Python 3.12, venv, PyTorch GPU wheels, TensorFlow, DeepFace, shell scripts
---
### Task 1: Add native deployment files
**Files:**
- Create: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/configs/native_x86_config.yaml`
- Create: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/run_rtsp.sh`
- Create: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/setup_native_venv.sh`
- Create: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/README_NATIVE.md`
**Step 1: Add a native x86 config**
- Point YOLO to the local project weight path.
- Keep RTSP behavior aligned with the current project.
**Step 2: Add a single editable RTSP launcher**
- Put `RTSP_URL` and `OUTPUT_DIR` at the top of the file.
- Run the project with `.venv/bin/python`.
**Step 3: Add a setup script**
- Create `.venv`
- Install GPU-enabled PyTorch
- Install TensorFlow and project requirements
### Task 2: Deploy to the target host
**Files:**
- No code changes required
**Step 1: Sync the updated project**
- Replace the target project directory while preserving weights if needed.
**Step 2: Ensure weights are in project-relative paths**
- Verify YOLO and DeepFace weights under `weights/`.
### Task 3: Validate and bundle
**Files:**
- No code changes required
**Step 1: Run setup on the target host**
- Execute the setup script.
**Step 2: Validate the RTSP CLI**
- Run `./.venv/bin/python main.py rtsp --help`.
**Step 3: Create the lightweight tarball**
- Exclude `.venv`
- Keep code, scripts, configs, docs, and weights

View File

@@ -0,0 +1,46 @@
# Offline Wheelhouse Design
**Date:** 2026-04-08
**Goal:** Add an offline Python dependency bundle for Ubuntu 24.04 x86_64 with Python 3.12 and NVIDIA GPU support so the project can be installed on similar machines without re-downloading PyTorch, TensorFlow, and application wheels.
## Scope
- Target platform: Ubuntu 24.04 x86_64
- Python version: 3.12
- GPU runtime: NVIDIA, using CUDA-enabled PyTorch wheels
- Bundle type: project code + weights + `wheelhouse/`
- Setup behavior: prefer offline wheels when present, fall back to network otherwise
## Approach
Add a dedicated wheelhouse build script that downloads:
- `pip`, `setuptools`, `wheel`
- `numpy<2`
- CUDA-enabled `torch` and `torchvision`
- `tensorflow[and-cuda]==2.16.1`
- `tf-keras==2.16.0`
- project requirements and their transitive dependencies
Store the wheels inside `wheelhouse/` under the project root.
Update the native setup script so it:
1. creates `.venv`
2. upgrades installer tooling from `wheelhouse/` when available
3. installs PyTorch and TensorFlow from local wheels when available
4. installs project requirements from local wheels when available
5. falls back to online indexes only if the wheelhouse is missing
## Bundle Layout
- `weights/`
- `wheelhouse/`
- `setup_native_venv.sh`
- `build_wheelhouse.sh`
- `run_rtsp.sh`
## Tradeoff
This increases the lightweight bundle size, but it removes repeat dependency downloads on future hosts. The user explicitly asked for an offline dependency pack, so this is the right tradeoff now.

View File

@@ -0,0 +1,51 @@
# Offline Wheelhouse Implementation Plan
> **For Claude:** REQUIRED SUB-SKILL: Use superpowers:executing-plans to implement this plan task-by-task.
**Goal:** Add a reusable offline wheelhouse for the native x86 bundle and make setup prefer local wheels.
**Architecture:** Keep the native bundle layout, add one build script that downloads all required wheels into `wheelhouse/`, and update the setup script to install from `wheelhouse/` first.
**Tech Stack:** Python 3.12, pip download, wheelhouse, shell scripts
---
### Task 1: Add offline dependency metadata and scripts
**Files:**
- Create: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/requirements-native.txt`
- Create: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/build_wheelhouse.sh`
- Modify: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/setup_native_venv.sh`
- Modify: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/README_NATIVE.md`
**Step 1: Add a native requirements file**
- Pin `numpy<2`
- Include app-level dependencies used by native setup
**Step 2: Add a wheelhouse build script**
- Download installer tools, PyTorch CUDA wheels, TensorFlow wheels, and project wheels
- Write everything into `wheelhouse/`
**Step 3: Make setup prefer offline wheels**
- Use `--no-index --find-links wheelhouse` when local wheels are available
- Fall back to online install otherwise
### Task 2: Sync and build the wheelhouse on the target host
**Files:**
- No code changes required
**Step 1: Sync project changes to `192.168.5.154`**
- Preserve existing weights
**Step 2: Run `build_wheelhouse.sh`**
- Populate `/home/x/people/people_flow_project/wheelhouse`
**Step 3: Validate setup behavior**
- Confirm `setup_native_venv.sh` recognizes local wheelhouse

View File

@@ -0,0 +1,31 @@
# RTSP Heartbeat Logging Design
**Date:** 2026-04-08
**Goal:** Add periodic heartbeat logs to the RTSP pipeline so operators can confirm the stream is still being processed during long 30-minute windows.
## Scope
- Keep the existing RTSP counting behavior unchanged.
- Print one heartbeat line every 60 seconds while the RTSP loop is running.
- Include the current demographic counts in the heartbeat output.
- Do not change JSON payload structure or window timing.
## Heartbeat Format
Each heartbeat line should report:
- runtime seconds
- current window index
- current window frame count
- total people in the active window
- age counts
- gender counts
- unknown attributes
- last processed timestamp
This output is intended for `tail -f` style monitoring and should remain single-line and compact.
## Approach
Reuse the existing live stats helper to avoid recomputing counting rules in a second place. The RTSP loop already knows when each sampled frame is processed, so it can track the last successful processing timestamp and emit a heartbeat when 60 seconds have elapsed since the last log.

View File

@@ -0,0 +1,47 @@
# RTSP Heartbeat Logging Implementation Plan
> **For Claude:** REQUIRED SUB-SKILL: Use superpowers:executing-plans to implement this plan task-by-task.
**Goal:** Add one-line RTSP heartbeat logs every 60 seconds so operators can monitor progress during long windows.
**Architecture:** Extend the RTSP loop with lightweight heartbeat state. Reuse the existing live stats builder and print one compact log line every 60 seconds after sampled frames are processed.
**Tech Stack:** Python, dataclasses, OpenCV, existing people-flow pipeline
---
### Task 1: Add heartbeat state and log output
**Files:**
- Modify: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/src/people_flow/pipeline.py`
**Step 1: Track heartbeat timing**
- Store the process start time.
- Store the next heartbeat deadline.
- Store the last successful processed timestamp.
**Step 2: Print one-line heartbeat logs**
- Reuse current live stats.
- Include runtime, window index, frame count, totals, demographics, unknown count, and last processed timestamp.
**Step 3: Keep the logging cadence stable**
- Emit at most one heartbeat per 60 seconds.
- Do not log on every frame.
### Task 2: Validate and synchronize
**Files:**
- No additional files required
**Step 1: Run compile checks**
Run: `python3 -m compileall main.py src`
Expected: PASS
**Step 2: Sync to remote host**
- Replace the remote project with the updated local copy.
- Keep the existing remote backup intact.

View File

@@ -0,0 +1,116 @@
# RTSP Windowed People Flow Design
**Date:** 2026-04-08
**Goal:** Extend the existing people-flow project with an RTSP mode that samples one frame per second from a live stream, computes people-flow and demographics, and writes a JSON summary every 30 minutes while preserving the existing offline video and batch modes.
## Scope
- Keep the existing `video` and `batch` commands unchanged.
- Add a new `rtsp` command for continuous live-stream processing.
- Sample one frame per second based on wall-clock time instead of processing every decoded frame.
- Maintain a 30-minute independent counting window.
- Write one timestamped JSON file per finished window.
- Refresh a `latest.json` file on every window flush.
- Do not save annotated RTSP video by default.
- Back up the current project before implementation.
## Approach
The current codebase already has reusable counting and attribute aggregation logic. The least risky change is to keep the offline pipeline as-is and add a dedicated RTSP processing path that reuses the same `LineCrossCounter` and `AttributeAggregator` components.
The RTSP path will:
1. Open an RTSP stream with OpenCV.
2. Read frames continuously.
3. Run inference only when at least one second has elapsed since the last processed frame.
4. Accumulate counts inside the current 30-minute window.
5. Flush a window summary to JSON when the window boundary is reached.
6. Reset all per-window state and continue into the next window.
7. Retry the stream connection when the RTSP source drops.
## Data Flow
### Command Layer
- `main.py` adds an `rtsp` subcommand with an `--input` RTSP URL.
- Existing global arguments such as `--config`, `--output-dir`, `--line`, and `--device` remain shared.
- RTSP mode disables video writing by default unless explicitly enabled in config later.
### Configuration
Add a new RTSP config section with:
- `sample_interval_seconds`
- `window_seconds`
- `reconnect_delay_seconds`
- `stream_open_timeout_seconds`
- `idle_sleep_seconds`
- `output_subdir`
This keeps timing and output behavior configurable without changing code.
### Processing Loop
Each processed frame will:
1. Pass through YOLO tracking.
2. Extract `person` track observations.
3. Optionally run DeepFace sampling on eligible tracks.
4. Update the line-cross counter.
5. Check whether the active 30-minute window should be flushed.
Skipped frames are decoded only to keep the stream current; they do not go through YOLO or DeepFace.
### Window Boundaries
Each window starts when the RTSP pipeline starts or right after the previous flush. The summary payload includes:
- `source_type`
- `source`
- `window_index`
- `window_start`
- `window_end`
- `window_duration_seconds`
- `total_people`
- `age_counts`
- `gender_counts`
- `unknown_attributes`
- `tracks`
After flushing:
- The timestamped JSON is written under `windows/`.
- `latest.json` is overwritten with the same payload.
- The counting and attribute state is reset.
## Output Layout
For `--output-dir /path/output`, the RTSP outputs live under:
- `/path/output/rtsp_stream/`
- `/path/output/rtsp_stream/latest.json`
- `/path/output/rtsp_stream/windows/stats_YYYY-MM-DD_HH-MM-SS.json`
The timestamp in the filename is the window end time.
## Error Handling
- If the RTSP stream cannot be opened, retry after a configurable delay.
- If frame reads fail mid-stream, release the capture and reconnect.
- If DeepFace analysis fails on a crop, treat that sample as unknown and keep running.
- If a window has zero crossings, still write a valid JSON payload with zero counts so downstream consumers can distinguish inactivity from pipeline failure.
## Compatibility
- `video` mode still writes annotated video and a final JSON after full processing.
- `batch` mode still writes a final CSV summary.
- Existing config keys remain valid.
## Testing Strategy
- Validate CLI parsing for the new `rtsp` command.
- Validate config loading with the new RTSP section.
- Validate that RTSP mode writes windowed JSON payloads and refreshes `latest.json`.
- Validate that 30-minute windows reset counts instead of accumulating indefinitely.
- Keep offline mode behavior intact by running `--help` and Python compile checks.

View File

@@ -0,0 +1,131 @@
# RTSP Windowed People Flow Implementation Plan
> **For Claude:** REQUIRED SUB-SKILL: Use superpowers:executing-plans to implement this plan task-by-task.
**Goal:** Add an RTSP mode that samples one frame per second, emits independent 30-minute JSON summaries, and preserves the existing offline video and batch workflows.
**Architecture:** Keep the existing offline pipeline untouched and add a dedicated RTSP pipeline path that reuses the counting and attribute aggregation components. Introduce a small RTSP configuration model and window-summary writer so the stream loop can reconnect, flush windowed JSON files, and reset state cleanly.
**Tech Stack:** Python, OpenCV, Ultralytics YOLO, DeepFace, PyYAML, dataclasses
---
### Task 1: Add RTSP configuration models
**Files:**
- Modify: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/src/people_flow/models.py`
- Modify: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/src/people_flow/config.py`
- Modify: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/configs/default_config.yaml`
**Step 1: Add an RTSP config dataclass**
- Add a dataclass with interval, window duration, reconnect delay, idle sleep, and output subdirectory fields.
- Attach it to `AppConfig`.
**Step 2: Load RTSP config from YAML**
- Update config loading to parse the new section.
- Keep backward compatibility when the section is absent.
**Step 3: Set sensible defaults in YAML**
- Add `sample_interval_seconds: 1`
- Add `window_seconds: 1800`
- Add reconnect and idle sleep defaults.
**Step 4: Run a compile check**
Run: `python3 -m compileall main.py src`
Expected: PASS
### Task 2: Add the RTSP CLI entrypoint
**Files:**
- Modify: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/main.py`
**Step 1: Add a new `rtsp` subcommand**
- Accept `--input` as the RTSP URL.
- Reuse global config and output arguments.
**Step 2: Wire the command to the pipeline**
- Call a new `process_rtsp()` method.
- Print the output directory and latest JSON path once the command starts.
**Step 3: Verify CLI help**
Run: `python3 main.py rtsp --help`
Expected: PASS and shows the RTSP input argument.
### Task 3: Implement the RTSP processing loop
**Files:**
- Modify: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/src/people_flow/pipeline.py`
- Modify: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/src/people_flow/io_utils.py`
**Step 1: Add RTSP output helpers**
- Add a helper that creates `/rtsp_stream/windows`.
- Add a helper that writes a timestamped JSON file and refreshes `latest.json`.
**Step 2: Add RTSP window summary generation**
- Reuse the existing summary-building logic, but parameterize it with `source`, `window_start`, and `window_end`.
- Keep the same count keys and track payload structure.
**Step 3: Add `process_rtsp()`**
- Open the RTSP stream with OpenCV.
- Reconnect on open/read failures after a delay.
- Sample one frame per second based on wall-clock time.
- Reuse YOLO tracking, crossing detection, and DeepFace aggregation on sampled frames only.
- Flush a JSON summary every 30 minutes.
- Reset counting and attribute state after each flush.
**Step 4: Keep long-running behavior explicit**
- Do not save annotated RTSP video by default.
- Ensure zero-count windows still emit JSON.
### Task 4: Preserve offline behavior
**Files:**
- Modify: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/src/people_flow/pipeline.py`
**Step 1: Refactor only shared summary code**
- Extract helper methods where useful.
- Do not change the existing `video`/`batch` outputs or file naming.
**Step 2: Re-run offline CLI smoke tests**
Run: `python3 main.py --help`
Expected: PASS
Run: `python3 main.py video --help`
Expected: PASS
Run: `python3 main.py batch --help`
Expected: PASS
### Task 5: Update docs and validate
**Files:**
- Modify: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/README.md`
**Step 1: Document the new RTSP mode**
- Add example commands.
- Explain the 1 FPS sampling and 30-minute window JSON behavior.
**Step 2: Run final validation**
Run: `python3 -m compileall main.py src`
Expected: PASS
Run: `python3 main.py rtsp --help`
Expected: PASS
Run: `python3 main.py --help`
Expected: PASS

View File

@@ -0,0 +1,79 @@
# x86 Docker Migration Design
**Date:** 2026-04-08
**Goal:** Package the RTSP people-flow project for direct use on an Ubuntu 24.04 x86_64 host with an NVIDIA RTX 3080 by using Docker, bundled project files, and host-side model weights.
## Scope
- Target host: `xiaozheng@192.168.5.154`
- Target path: `/home/x/people`
- Runtime model: Docker with NVIDIA runtime
- Input source: RTSP
- Output: JSON window summaries under a mounted host directory
- Include required model weights on the target host
## Why Docker
The existing remote runtime was built on Jetson ARM64 and cannot be reused on an x86_64 RTX 3080 machine. The target host only has Python 3.12 installed, and a native port would need additional interpreter and CUDA-specific package work. Docker is the most reliable path because it isolates Python dependencies, preserves a reproducible runtime, and matches the user requirement of direct use on a new CUDA-capable machine.
## Packaging Strategy
### Host Layout
The target host will contain:
- `/home/x/people/people_flow_project/`
- `/home/x/people/people_flow_project/weights/yolo11n.pt`
- `/home/x/people/people_flow_project/weights/deepface/age_model_weights.h5`
- `/home/x/people/people_flow_project/weights/deepface/gender_model_weights.h5`
- `/home/x/people/people_flow_project/weights/deepface/retinaface.h5`
- `/home/x/people/output/`
### Container Layout
The container will:
- run on Python 3.12
- install GPU-enabled PyTorch wheels
- install the application dependencies
- read YOLO and DeepFace weights from deterministic in-container paths
- write outputs to a mounted host output directory
The project source will be copied into the image at build time. The host-side `weights/` directory will also be part of the build context so the final image does not need to download weights on first start.
## Runtime Contract
The image is intended to be built once on the target host and then started with a single `docker run` command using `--gpus all`.
The container command will remain the existing CLI:
`python main.py --config ... --output-dir ... --device cuda:0 rtsp --input ...`
## System Adaptation
The target host already has:
- Ubuntu 24.04
- Docker installed
- NVIDIA runtime registered in Docker
The adaptation work is therefore limited to:
- adding the projects Docker packaging files
- transferring project code and model weights
- building the image on the target host
- validating the container entrypoint and GPU runtime path
## Risks
- The target GPU is currently heavily occupied by another process, so a full inference validation may need to avoid competing for memory.
- DeepFace and TensorFlow increase image size and build time.
- Network access is required during image build unless a wheel cache is prepared separately.
## Success Criteria
- The target host contains the project and all required weights under `/home/x/people`
- `docker build` completes successfully
- The container can run `main.py rtsp --help`
- The final run command is documented for direct RTSP use

View File

@@ -0,0 +1,77 @@
# x86 Docker Migration Implementation Plan
> **For Claude:** REQUIRED SUB-SKILL: Use superpowers:executing-plans to implement this plan task-by-task.
**Goal:** Make the RTSP people-flow project directly usable on Ubuntu 24.04 x86_64 with an RTX 3080 by transferring code and weights and building a Docker image on the target host.
**Architecture:** Use a Docker-based runtime for Python 3.12, GPU-enabled PyTorch, DeepFace, and the existing project CLI. Keep weights in a deterministic project directory and bake them into the image during build so runtime startup does not trigger downloads.
**Tech Stack:** Docker, NVIDIA Container Runtime, Python 3.12, PyTorch, Ultralytics, DeepFace
---
### Task 1: Add Docker packaging files
**Files:**
- Create: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/Dockerfile`
- Create: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/docker-compose.yml`
- Create: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/.dockerignore`
- Create: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/scripts/run_rtsp_docker.sh`
- Modify: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/README.md`
**Step 1: Define the image build**
- Base the image on Python 3.12.
- Install required OS packages for OpenCV and ffmpeg.
- Install GPU-enabled PyTorch and project dependencies.
- Copy project source and weights into the image.
**Step 2: Add a Docker run wrapper**
- Provide a shell script that accepts RTSP URL and output directory.
- Use `--gpus all`.
**Step 3: Update the README**
- Document the Docker build and run commands.
- Document where weights must live if the host directory is rebuilt.
### Task 2: Prepare the project tree for weights
**Files:**
- Create: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/weights/.gitkeep`
- Create: `/Users/zxmacmini1/Documents/人流检测/people_flow_project/weights/deepface/.gitkeep`
**Step 1: Create weight directories**
- Reserve stable paths for YOLO and DeepFace weights.
### Task 3: Transfer the project and weights to the target host
**Files:**
- No code changes required
**Step 1: Copy the project**
- Transfer the project directory to `/home/x/people/people_flow_project`.
**Step 2: Copy YOLO and DeepFace weights**
- Place YOLO and DeepFace weights into the target project `weights/` tree.
### Task 4: Build and validate on the target host
**Files:**
- No code changes required
**Step 1: Build the image**
- Run `docker build` under `/home/x/people/people_flow_project`.
**Step 2: Validate the CLI**
- Run the container with `python main.py rtsp --help`.
**Step 3: Provide the final RTSP run command**
- Document the exact `docker run` invocation for the target host.

View File

@@ -0,0 +1,136 @@
from __future__ import annotations
import argparse
from pathlib import Path
def build_parser() -> argparse.ArgumentParser:
parser = argparse.ArgumentParser(
description="People-flow counting with YOLO tracking and DeepFace demographics."
)
parser.add_argument(
"--config",
default="configs/default_config.yaml",
help="Path to the YAML config file.",
)
parser.add_argument(
"--output-dir",
default=None,
help="Directory for generated artifacts.",
)
parser.add_argument(
"--line",
help="Override counting line as x1,y1,x2,y2.",
)
parser.add_argument(
"--line-mode",
choices=["normalized", "pixel"],
help="Coordinate mode for --line.",
)
parser.add_argument(
"--device",
help="Override inference device, for example cuda:0 or cpu.",
)
subparsers = parser.add_subparsers(dest="command", required=True)
video_parser = subparsers.add_parser("video", help="Process one video.")
video_parser.add_argument("--input", required=True, help="Path to the video file.")
video_parser.add_argument(
"--skip-video-save",
action="store_true",
help="Do not write the annotated video.",
)
batch_parser = subparsers.add_parser("batch", help="Process a directory of videos.")
batch_parser.add_argument(
"--input-dir",
required=True,
help="Directory scanned recursively for videos.",
)
batch_parser.add_argument(
"--pattern",
default="*.mp4",
help="Glob pattern used during recursive discovery.",
)
batch_parser.add_argument(
"--skip-video-save",
action="store_true",
help="Do not write annotated videos.",
)
rtsp_parser = subparsers.add_parser("rtsp", help="Process a live RTSP stream.")
rtsp_parser.add_argument("--input", help="RTSP URL.")
manage_api_parser = subparsers.add_parser("manage-api", help="Start the management API.")
manage_api_parser.add_argument("--host", default="0.0.0.0", help="Host for the management API.")
manage_api_parser.add_argument("--port", type=int, default=18082, help="Port for the management API.")
return parser
def build_config(args: argparse.Namespace):
from src.people_flow.config import load_config, merge_cli_overrides
save_video = None
if hasattr(args, "skip_video_save"):
save_video = not args.skip_video_save
config = load_config(Path(args.config))
return merge_cli_overrides(
config=config,
line=args.line,
line_mode=args.line_mode,
device=args.device,
save_video=save_video,
)
def main() -> int:
parser = build_parser()
args = parser.parse_args()
if args.command == "manage-api":
from src.people_flow.manage_api import run_manage_api
run_manage_api(args.config, host=args.host, port=args.port)
return 0
config = build_config(args)
from src.people_flow.pipeline import PeopleFlowPipeline, discover_videos
output_root = Path(args.output_dir or config.runtime.output_dir)
pipeline = PeopleFlowPipeline(config=config, output_root=output_root)
if args.command == "rtsp":
paths = pipeline.get_rtsp_output_paths()
print(f"rtsp_output_dir={paths['root']}", flush=True)
print(f"latest_json={paths['latest_json']}", flush=True)
source = args.input or config.runtime.rtsp_url
if not source:
raise SystemExit("RTSP source is required. Pass --input or set runtime.rtsp_url in the config.")
pipeline.process_rtsp(source)
return 0
if args.command == "video":
result = pipeline.process_video(Path(args.input))
print(f"processed_video={result['video_name']}")
print(f"total_people={result['total_people']}")
print(f"unknown_attributes={result['unknown_attributes']}")
print(f"json={result['json_path']}")
if result.get("video_output_path"):
print(f"annotated_video={result['video_output_path']}")
return 0
videos = discover_videos(Path(args.input_dir), pattern=args.pattern)
if not videos:
raise SystemExit(f"No videos found under {args.input_dir} with pattern {args.pattern}")
summary = pipeline.process_batch(videos)
print(f"videos_processed={len(summary['videos'])}")
print(f"csv={summary['csv_path']}")
return 0
if __name__ == "__main__":
raise SystemExit(main())

View File

@@ -0,0 +1,11 @@
[project]
name = "people-flow-project"
version = "0.1.0"
description = "Street video people-flow counting with YOLO tracking and face-based age/gender estimation"
readme = "README.md"
requires-python = ">=3.10,<3.13"
dependencies = []
[build-system]
requires = ["setuptools>=68"]
build-backend = "setuptools.build_meta"

View File

@@ -0,0 +1,7 @@
flask>=3.1.0
ultralytics>=8.3.0
opencv-python-headless>=4.10.0
deepface>=0.0.93
pyyaml>=6.0.2
pandas>=2.2.3
numpy<2

View File

@@ -0,0 +1,8 @@
flask>=3.1.0
numpy<2
ultralytics==8.4.35
lap>=0.5.12
opencv-python==4.11.0.86
deepface==0.0.99
pyyaml==6.0.3
pandas==3.0.2

View File

@@ -0,0 +1,6 @@
ultralytics>=8.3.0
opencv-python>=4.10.0
deepface>=0.0.93
pyyaml>=6.0.2
pandas>=2.2.3
numpy>=1.26.0

View File

@@ -0,0 +1,6 @@
#!/usr/bin/env bash
set -euo pipefail
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
exec "$SCRIPT_DIR/scripts/run.sh" "$@"

View File

@@ -0,0 +1,40 @@
#!/usr/bin/env sh
set -eu
PROJECT_DIR="/opt/people-flow"
CONFIG_TEMPLATE="${PROJECT_DIR}/config/config.example.yaml"
CONFIG_PATH="${CONFIG_PATH:-${PROJECT_DIR}/config/local.yaml}"
OUTPUT_DIR="${OUTPUT_DIR:-${PROJECT_DIR}/outputs}"
RTSP_URL="${RTSP_URL:-}"
API_HOST="${API_HOST:-0.0.0.0}"
API_PORT="${API_PORT:-18082}"
mkdir -p "${OUTPUT_DIR}" "$(dirname "${CONFIG_PATH}")"
if [ ! -f "${CONFIG_PATH}" ]; then
cp "${CONFIG_TEMPLATE}" "${CONFIG_PATH}"
fi
python - "$CONFIG_PATH" "$RTSP_URL" "$OUTPUT_DIR" <<'PY'
from pathlib import Path
import sys
import yaml
config_path = Path(sys.argv[1])
rtsp_url = sys.argv[2]
output_dir = sys.argv[3]
raw = yaml.safe_load(config_path.read_text(encoding="utf-8")) or {}
runtime = raw.setdefault("runtime", {})
if rtsp_url:
runtime["rtsp_url"] = rtsp_url
runtime["output_dir"] = output_dir
yolo = raw.setdefault("yolo", {})
yolo.setdefault("model_path", "weights/yolo11n.pt")
config_path.write_text(
yaml.safe_dump(raw, allow_unicode=True, sort_keys=False),
encoding="utf-8",
)
PY
exec python main.py --config "${CONFIG_PATH}" manage-api --host "${API_HOST}" --port "${API_PORT}"

View File

@@ -0,0 +1,60 @@
#!/usr/bin/env bash
set -euo pipefail
PROJECT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
SETUP_SCRIPT="${PROJECT_DIR}/setup_native_venv.sh"
RUN_SCRIPT="${PROJECT_DIR}/scripts/run.sh"
INSTALL_SERVICE_SCRIPT="${PROJECT_DIR}/scripts/install_service.sh"
PROJECT_USER="${SUDO_USER:-$(id -un)}"
run_privileged() {
if [[ "$(id -u)" -eq 0 ]]; then
"$@"
return
fi
sudo "$@"
}
run_project_user() {
if [[ "$(id -u)" -eq 0 && -n "${SUDO_USER:-}" ]]; then
sudo -u "${PROJECT_USER}" -H "$@"
return
fi
"$@"
}
ensure_system_package() {
local command_name="$1"
local package_name="$2"
if command -v "${command_name}" >/dev/null 2>&1; then
return
fi
echo "Installing missing package: ${package_name}"
run_privileged apt-get -o Acquire::ForceIPv4=true update
run_privileged apt-get -o Acquire::ForceIPv4=true install -y "${package_name}"
}
ensure_system_package ffmpeg ffmpeg
if [[ ! -d "/usr/lib/python3.12/venv" && ! -d "/usr/lib/python3.12/ensurepip" ]]; then
echo "Installing missing package: python3.12-venv"
run_privileged apt-get -o Acquire::ForceIPv4=true update
run_privileged apt-get -o Acquire::ForceIPv4=true install -y python3.12-venv
fi
if ! command -v nvidia-smi >/dev/null 2>&1; then
echo "nvidia-smi is required but not installed." >&2
exit 1
fi
run_project_user env PYTHON_BIN="${PYTHON_BIN:-python3.12}" bash "${SETUP_SCRIPT}"
run_project_user bash "${RUN_SCRIPT}" --prepare-only
bash "${INSTALL_SERVICE_SCRIPT}"
run_privileged systemctl enable --now people-flow.service
cat <<EOF
Offline install complete.
Service started and enabled on boot: people-flow.service
Runtime log: ${PROJECT_DIR}/outputs/rtsp_run.log
EOF

View File

@@ -0,0 +1,33 @@
#!/usr/bin/env bash
set -euo pipefail
PROJECT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
TEMPLATE_PATH="${PROJECT_DIR}/deploy/people-flow.service.tpl"
CONFIG_PATH="${CONFIG_PATH:-${PROJECT_DIR}/config/local.yaml}"
SERVICE_NAME="${SERVICE_NAME:-people-flow.service}"
OUTPUT_PATH="${PROJECT_DIR}/deploy/${SERVICE_NAME}"
RUN_USER="${RUN_USER:-${SUDO_USER:-$(id -un)}}"
RUN_GROUP="${RUN_GROUP:-$(id -gn "${RUN_USER}")}"
if [[ ! -f "${TEMPLATE_PATH}" ]]; then
echo "Missing service template: ${TEMPLATE_PATH}" >&2
exit 1
fi
if [[ ! -f "${CONFIG_PATH}" ]]; then
echo "Missing config file: ${CONFIG_PATH}" >&2
exit 1
fi
sed \
-e "s|__PROJECT_DIR__|${PROJECT_DIR}|g" \
-e "s|__CONFIG_PATH__|${CONFIG_PATH}|g" \
-e "s|__RUN_USER__|${RUN_USER}|g" \
-e "s|__RUN_GROUP__|${RUN_GROUP}|g" \
"${TEMPLATE_PATH}" > "${OUTPUT_PATH}"
sudo cp "${OUTPUT_PATH}" "/etc/systemd/system/${SERVICE_NAME}"
sudo systemctl daemon-reload
echo "Service installed to /etc/systemd/system/${SERVICE_NAME}"
echo "Enable and start it with: sudo systemctl enable --now ${SERVICE_NAME}"

View File

@@ -0,0 +1,41 @@
#!/usr/bin/env bash
set -euo pipefail
PROJECT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
VENV_PYTHON="${PROJECT_DIR}/.venv/bin/python"
CONFIG_TEMPLATE="${PROJECT_DIR}/config/config.example.yaml"
CONFIG_PATH="${PROJECT_DIR}/config/local.yaml"
RTSP_URL="${RTSP_URL:-rtsp://user:password@camera-ip:554/h264/ch1/main/av_stream}"
OUTPUT_DIR="${OUTPUT_DIR:-${PROJECT_DIR}/outputs}"
PREPARE_ONLY=0
if [[ "${1:-}" == "--prepare-only" ]]; then
PREPARE_ONLY=1
shift
fi
if [[ ! -x "${VENV_PYTHON}" ]]; then
echo "Virtual environment is missing. Run scripts/install.sh first." >&2
exit 1
fi
if [[ "${RTSP_URL}" == "rtsp://user:password@camera-ip:554/h264/ch1/main/av_stream" ]]; then
echo "Please edit scripts/run.sh and set RTSP_URL before starting." >&2
exit 1
fi
mkdir -p "${OUTPUT_DIR}" "${PROJECT_DIR}/config"
cp "${CONFIG_TEMPLATE}" "${CONFIG_PATH}"
sed -i.bak \
-e "s|^ rtsp_url: .*| rtsp_url: \"${RTSP_URL}\"|" \
-e "s|^ output_dir: .*| output_dir: \"${OUTPUT_DIR}\"|" \
"${CONFIG_PATH}"
rm -f "${CONFIG_PATH}.bak"
if [[ "${PREPARE_ONLY}" -eq 1 ]]; then
echo "Prepared config at ${CONFIG_PATH}"
exit 0
fi
exec "${VENV_PYTHON}" "${PROJECT_DIR}/main.py" --config "${CONFIG_PATH}" rtsp "$@"

View File

@@ -0,0 +1,26 @@
#!/usr/bin/env bash
set -euo pipefail
if [[ $# -lt 2 ]]; then
echo "Usage: $0 <rtsp_url> <host_output_dir>"
exit 1
fi
RTSP_URL="$1"
HOST_OUTPUT_DIR="$2"
mkdir -p "$HOST_OUTPUT_DIR"
docker run -d \
--name people-flow-rtsp \
--restart unless-stopped \
--network host \
--gpus all \
--shm-size 1g \
-v "$HOST_OUTPUT_DIR:/opt/people-flow/output" \
people-flow-rtsp:x86-cuda \
--config /opt/people-flow/configs/docker_x86_config.yaml \
--output-dir /opt/people-flow/output \
--device cuda:0 \
rtsp \
--input "$RTSP_URL"

View File

@@ -0,0 +1,38 @@
#!/usr/bin/env bash
set -euo pipefail
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
PROJECT_ROOT="$SCRIPT_DIR"
WHEELHOUSE_DIR="$PROJECT_ROOT/wheelhouse"
DEEPFACE_SOURCE_DIR="$PROJECT_ROOT/weights/deepface"
DEEPFACE_TARGET_DIR="${HOME}/.deepface/weights"
PYTHON_BIN="${PYTHON_BIN:-python3.12}"
cd "$PROJECT_ROOT"
"$PYTHON_BIN" -m venv .venv
source .venv/bin/activate
if [[ -d "$WHEELHOUSE_DIR" ]] && find "$WHEELHOUSE_DIR" -maxdepth 1 -name '*.whl' | grep -q .; then
python -m pip install --no-index --find-links "$WHEELHOUSE_DIR" --upgrade pip setuptools wheel
pip install --no-index --find-links "$WHEELHOUSE_DIR" "numpy<2"
pip install --no-index --find-links "$WHEELHOUSE_DIR" torch torchvision
pip install --no-index --find-links "$WHEELHOUSE_DIR" "tensorflow[and-cuda]==2.16.1" "tf-keras==2.16.0"
pip install --no-index --find-links "$WHEELHOUSE_DIR" -r requirements-native.txt
else
python -m pip install --upgrade pip setuptools wheel
pip install "numpy<2"
pip install --index-url https://download.pytorch.org/whl/cu126 torch torchvision
pip install "tensorflow[and-cuda]==2.16.1" "tf-keras==2.16.0"
pip install -r requirements-native.txt
fi
mkdir -p "$DEEPFACE_TARGET_DIR"
if find "$DEEPFACE_SOURCE_DIR" -maxdepth 1 -name '*.h5' | grep -q .; then
cp -f "$DEEPFACE_SOURCE_DIR"/*.h5 "$DEEPFACE_TARGET_DIR"/
else
echo "Warning: missing bundled DeepFace weights under $DEEPFACE_SOURCE_DIR"
echo "Attribute analysis will stay unavailable until the .h5 files are provided."
fi
echo "venv_ready=$PROJECT_ROOT/.venv"

View File

@@ -0,0 +1 @@
"""People flow analysis package."""

View File

@@ -0,0 +1,141 @@
from __future__ import annotations
from collections import Counter, defaultdict
from statistics import median
from typing import Any
import cv2
import numpy as np
from .models import AttributeConfig, AttributeVote, TrackAttributeSummary, TrackObservation
def age_to_bucket(age: int) -> str:
if age < 18:
return "minor"
if age < 60:
return "adult"
return "senior"
def normalize_gender(raw_gender: str | None) -> str | None:
if not raw_gender:
return None
lowered = raw_gender.strip().lower()
if lowered in {"man", "male"}:
return "male"
if lowered in {"woman", "female"}:
return "female"
return None
class AttributeAggregator:
def __init__(self, config: AttributeConfig) -> None:
self.config = config
self.votes: dict[int, list[AttributeVote]] = defaultdict(list)
self.samples_taken: dict[int, int] = defaultdict(int)
self.last_sampled_frame: dict[int, int] = {}
self._deepface = self._load_deepface() if config.enabled else None
def _load_deepface(self) -> Any:
try:
from deepface import DeepFace
except ImportError as exc:
raise RuntimeError(
"DeepFace is not installed. Install dependencies with `pip install -r requirements.txt`."
) from exc
return DeepFace
def maybe_collect(self, frame: np.ndarray, frame_index: int, track: TrackObservation) -> None:
if self._deepface is None:
return
if self.samples_taken[track.track_id] >= self.config.max_samples_per_track:
return
last_frame = self.last_sampled_frame.get(track.track_id)
if last_frame is not None and frame_index - last_frame < self.config.sample_every_n_frames:
return
x1, y1, x2, y2 = track.bbox
width = x2 - x1
height = y2 - y1
if width < self.config.min_person_box_width or height < self.config.min_person_box_height:
return
crop = self._crop_person(frame, track.bbox)
if crop.size == 0:
return
vote = self._analyze_crop(crop)
self.last_sampled_frame[track.track_id] = frame_index
if vote is None:
return
self.samples_taken[track.track_id] += 1
self.votes[track.track_id].append(vote)
def reset(self) -> None:
self.votes.clear()
self.samples_taken.clear()
self.last_sampled_frame.clear()
def _crop_person(self, frame: np.ndarray, bbox: tuple[int, int, int, int]) -> np.ndarray:
x1, y1, x2, y2 = bbox
height, width = frame.shape[:2]
pad_x = int((x2 - x1) * self.config.person_crop_padding)
pad_y = int((y2 - y1) * self.config.person_crop_padding)
left = max(0, x1 - pad_x)
top = max(0, y1 - pad_y)
right = min(width, x2 + pad_x)
bottom = min(height, y2 + pad_y)
return frame[top:bottom, left:right]
def _analyze_crop(self, crop: np.ndarray) -> AttributeVote | None:
rgb_crop = cv2.cvtColor(crop, cv2.COLOR_BGR2RGB)
try:
analysis = self._deepface.analyze(
img_path=rgb_crop,
actions=["age", "gender"],
detector_backend=self.config.detector_backend,
enforce_detection=self.config.enforce_detection,
silent=True,
)
except Exception:
return None
if isinstance(analysis, list):
if not analysis:
return None
analysis = analysis[0]
age_value = analysis.get("age")
gender_value = normalize_gender(analysis.get("dominant_gender"))
if age_value is None or gender_value is None:
return None
age_int = int(round(float(age_value)))
return AttributeVote(
age=age_int,
age_bucket=age_to_bucket(age_int),
gender=gender_value,
)
def summarize_track(self, track_id: int) -> TrackAttributeSummary | None:
votes = self.votes.get(track_id, [])
if not votes:
return None
age_bucket_counts = Counter(vote.age_bucket for vote in votes)
gender_counts = Counter(vote.gender for vote in votes)
if not age_bucket_counts or not gender_counts:
return None
age_bucket = age_bucket_counts.most_common(1)[0][0]
gender = gender_counts.most_common(1)[0][0]
age_value = int(round(median(vote.age for vote in votes)))
return TrackAttributeSummary(
track_id=track_id,
age=age_value,
age_bucket=age_bucket,
gender=gender,
samples_used=len(votes),
)

View File

@@ -0,0 +1,99 @@
from __future__ import annotations
from dataclasses import replace
from pathlib import Path
import yaml
from .models import (
AppConfig,
AttributeConfig,
CountingConfig,
OutputConfig,
RtspConfig,
RuntimeConfig,
YoloConfig,
)
def _read_yaml(config_path: Path) -> dict:
if not config_path.exists():
raise FileNotFoundError(f"Config file not found: {config_path}")
with config_path.open("r", encoding="utf-8") as handle:
loaded = yaml.safe_load(handle) or {}
if not isinstance(loaded, dict):
raise ValueError(f"Config file must contain a mapping: {config_path}")
return loaded
def load_config_document(config_path: Path) -> dict:
return _read_yaml(config_path)
def save_config_document(config_path: Path, payload: dict) -> None:
config_path.parent.mkdir(parents=True, exist_ok=True)
temp_path = config_path.with_suffix(config_path.suffix + ".tmp")
temp_path.write_text(
yaml.safe_dump(payload, allow_unicode=True, sort_keys=False),
encoding="utf-8",
)
temp_path.replace(config_path)
def resolve_project_root(config_path: Path) -> Path:
return config_path.expanduser().resolve().parent.parent
def resolve_project_path(project_root: Path, raw_path: str | Path) -> Path:
path = Path(raw_path)
if path.is_absolute():
return path.resolve()
return (project_root.resolve() / path).resolve()
def load_config(config_path: Path) -> AppConfig:
data = _read_yaml(config_path)
config = AppConfig(
yolo=YoloConfig(**data.get("yolo", {})),
counting=CountingConfig(**_normalize_counting_config(data.get("counting", {}))),
attributes=AttributeConfig(**data.get("attributes", {})),
output=OutputConfig(**data.get("output", {})),
rtsp=RtspConfig(**data.get("rtsp", {})),
runtime=RuntimeConfig(**data.get("runtime", {})),
config_path=config_path.resolve(),
)
return config
def _normalize_counting_config(data: dict) -> dict:
normalized = dict(data)
line = normalized.get("line")
if line is not None:
normalized["line"] = tuple(float(value) for value in line)
return normalized
def parse_line_override(raw_line: str) -> tuple[float, float, float, float]:
parts = [part.strip() for part in raw_line.split(",")]
if len(parts) != 4:
raise ValueError("--line must contain exactly four comma-separated values")
return tuple(float(part) for part in parts) # type: ignore[return-value]
def merge_cli_overrides(
config: AppConfig,
line: str | None,
line_mode: str | None,
device: str | None,
save_video: bool | None,
) -> AppConfig:
updated = config
if line:
updated.counting = replace(updated.counting, line=parse_line_override(line))
if line_mode:
updated.counting = replace(updated.counting, line_mode=line_mode)
if device:
updated.yolo = replace(updated.yolo, device=device)
if save_video is not None:
updated.output = replace(updated.output, save_video=save_video)
return updated

View File

@@ -0,0 +1,52 @@
from __future__ import annotations
from .models import CountingConfig, CrossingEvent, TrackObservation
def _line_side(
point: tuple[float, float], line: tuple[float, float, float, float]
) -> float:
px, py = point
x1, y1, x2, y2 = line
return (x2 - x1) * (py - y1) - (y2 - y1) * (px - x1)
class LineCrossCounter:
def __init__(self, line: tuple[float, float, float, float], config: CountingConfig) -> None:
self.line = line
self.config = config
self.previous_side: dict[int, float] = {}
self.counted_ids: set[int] = set()
self.crossings: list[CrossingEvent] = []
def update(self, observations: list[TrackObservation]) -> list[CrossingEvent]:
events: list[CrossingEvent] = []
for observation in observations:
side = _line_side(observation.center, self.line)
previous = self.previous_side.get(observation.track_id)
self.previous_side[observation.track_id] = side
if observation.track_id in self.counted_ids:
continue
if previous is None:
continue
if abs(previous) <= self.config.crossing_tolerance or abs(side) <= self.config.crossing_tolerance:
continue
if previous * side >= 0:
continue
direction = "negative_to_positive" if previous < 0 < side else "positive_to_negative"
event = CrossingEvent(track_id=observation.track_id, direction=direction)
self.counted_ids.add(observation.track_id)
self.crossings.append(event)
events.append(event)
return events
def reset(self) -> None:
self.previous_side.clear()
self.counted_ids.clear()
self.crossings.clear()
@property
def total_people(self) -> int:
return len(self.counted_ids)

View File

@@ -0,0 +1,95 @@
from __future__ import annotations
import json
from datetime import datetime
from pathlib import Path
import cv2
from .models import TrackObservation
def ensure_dir(path: Path) -> Path:
path.mkdir(parents=True, exist_ok=True)
return path
def make_video_writer(path: Path, width: int, height: int, fps: float) -> cv2.VideoWriter:
fourcc = cv2.VideoWriter_fourcc(*"mp4v")
return cv2.VideoWriter(str(path), fourcc, fps if fps > 0 else 25.0, (width, height))
def write_json(path: Path, payload: dict) -> None:
with path.open("w", encoding="utf-8") as handle:
json.dump(payload, handle, ensure_ascii=True, indent=2)
def write_csv(path: Path, rows: list[dict]) -> None:
import pandas as pd
dataframe = pd.DataFrame(rows)
dataframe.to_csv(path, index=False)
def write_window_json(windows_dir: Path, latest_path: Path, payload: dict, window_end: datetime) -> Path:
ensure_dir(windows_dir)
ensure_dir(latest_path.parent)
target = windows_dir / f"stats_{window_end.strftime('%Y-%m-%d_%H-%M-%S')}.json"
write_json(target, payload)
write_json(latest_path, payload)
return target
def draw_line(frame, line: tuple[float, float, float, float]) -> None:
x1, y1, x2, y2 = (int(value) for value in line)
cv2.line(frame, (x1, y1), (x2, y2), (0, 255, 255), 2)
def draw_tracks(
frame,
observations: list[TrackObservation],
counted_ids: set[int],
draw_labels: bool,
) -> None:
for observation in observations:
x1, y1, x2, y2 = observation.bbox
color = (0, 200, 0) if observation.track_id in counted_ids else (255, 140, 0)
cv2.rectangle(frame, (x1, y1), (x2, y2), color, 2)
if draw_labels:
label = f"id={observation.track_id} conf={observation.confidence:.2f}"
cv2.putText(
frame,
label,
(x1, max(20, y1 - 6)),
cv2.FONT_HERSHEY_SIMPLEX,
0.5,
color,
1,
cv2.LINE_AA,
)
def draw_stats(frame, stats: dict) -> None:
lines = [
f"total_people: {stats['total_people']}",
f"minor: {stats['age_counts']['minor']}",
f"adult: {stats['age_counts']['adult']}",
f"senior: {stats['age_counts']['senior']}",
f"male: {stats['gender_counts']['male']}",
f"female: {stats['gender_counts']['female']}",
f"unknown_attributes: {stats['unknown_attributes']}",
]
x = 12
y = 24
for text in lines:
cv2.putText(
frame,
text,
(x, y),
cv2.FONT_HERSHEY_SIMPLEX,
0.65,
(255, 255, 255),
2,
cv2.LINE_AA,
)
y += 24

View File

@@ -0,0 +1,389 @@
from __future__ import annotations
import json
from argparse import ArgumentParser
from dataclasses import dataclass
from datetime import datetime
from pathlib import Path
from flask import Flask, jsonify, request, send_file
from .config import (
load_config,
load_config_document,
resolve_project_path,
resolve_project_root,
save_config_document,
)
PROJECT_TYPE = "people_flow_project"
DEFAULT_MANAGE_PORT = 18082
MAX_PREVIEW_LINES = 2000
@dataclass(slots=True)
class ManageContext:
config_path: Path
project_root: Path
def create_app(config_path: str | Path) -> Flask:
resolved_config = Path(config_path).expanduser().resolve()
ctx = ManageContext(
config_path=resolved_config,
project_root=resolve_project_root(resolved_config),
)
app = Flask(__name__)
app.config["MANAGE_CONTEXT"] = ctx
@app.get("/api/manage/health")
def get_health():
return jsonify(
{
"status": "ok",
"project_type": PROJECT_TYPE,
"version": "dev",
"runtime_status": "running",
}
)
@app.get("/api/manage/config")
def get_config():
return jsonify(_config_payload(ctx))
@app.put("/api/manage/config")
def update_config():
payload = request.get_json(silent=True) or {}
rtsp_url = payload.get("rtsp_url")
if not isinstance(rtsp_url, str) or not rtsp_url.strip():
return jsonify({"error": "rtsp_url is required"}), 400
raw = load_config_document(ctx.config_path)
runtime = raw.setdefault("runtime", {})
runtime["rtsp_url"] = rtsp_url.strip()
save_config_document(ctx.config_path, raw)
return jsonify(_config_payload(ctx))
@app.get("/api/manage/summary")
def get_summary():
return jsonify(_build_summary(ctx))
@app.get("/api/manage/windows")
def get_windows():
page = max(_int_arg("page", 1), 1)
page_size = max(_int_arg("page_size", 24), 1)
limit = request.args.get("limit")
items = list(_load_window_stats(ctx))
if limit is not None:
items = items[: max(_int_value(limit), 0)]
start = (page - 1) * page_size
end = start + page_size
return jsonify(
{
"items": items[start:end],
"page": page,
"page_size": page_size,
"total": len(items),
}
)
@app.get("/api/manage/files")
def get_files():
return jsonify({"files": _list_result_files(ctx)})
@app.get("/api/manage/files/preview")
def preview_file():
target = _resolve_sandbox_file(ctx, request.args.get("path", ""))
lines = _tail_lines(target, _bounded_preview_lines(request.args.get("lines")))
return jsonify(
{
"path": _relative_path(ctx, target),
"lines": lines,
"count": len(lines),
}
)
@app.get("/api/manage/files/download")
def download_file():
target = _resolve_sandbox_file(ctx, request.args.get("path", ""))
return send_file(target, as_attachment=True, download_name=target.name)
@app.errorhandler(ValueError)
def handle_value_error(error: ValueError):
return jsonify({"error": str(error)}), 400
@app.errorhandler(FileNotFoundError)
def handle_missing_file(error: FileNotFoundError):
return jsonify({"error": str(error)}), 404
return app
def run_manage_api(
config_path: str | Path,
host: str = "0.0.0.0",
port: int = DEFAULT_MANAGE_PORT,
) -> None:
app = create_app(config_path)
app.run(host=host, port=port)
def parse_args() -> ArgumentParser:
parser = ArgumentParser(description="People flow management API")
parser.add_argument("--config", required=True, help="Path to YAML config file")
parser.add_argument("--host", default="0.0.0.0", help="Host for the management API")
parser.add_argument("--port", type=int, default=DEFAULT_MANAGE_PORT, help="Port for the management API")
return parser
def main() -> int:
parser = parse_args()
args = parser.parse_args()
run_manage_api(args.config, host=args.host, port=args.port)
return 0
def _config_payload(ctx: ManageContext) -> dict:
config = load_config(ctx.config_path)
output_root = resolve_project_path(ctx.project_root, config.runtime.output_dir)
return {
"project_type": PROJECT_TYPE,
"config_path": str(ctx.config_path),
"runtime": {
"rtsp_url": config.runtime.rtsp_url,
"output_dir": str(output_root),
},
"rtsp": {
"output_subdir": config.rtsp.output_subdir,
"window_seconds": config.rtsp.window_seconds,
},
}
def _build_summary(ctx: ManageContext) -> dict:
summary_path, payload = _load_summary_payload(ctx)
all_window_stats = _load_window_stats(ctx)
if payload is None:
latest_json = _latest_json_path(ctx)
return {
"result_type": PROJECT_TYPE,
"headline": "No RTSP summary output yet",
"metrics": {
"latest_path": str(latest_json),
"recent_window_stats": all_window_stats[:24],
"all_window_stats": all_window_stats,
},
}
tracks = payload.get("tracks", [])
direction_counts: dict[str, int] = {}
if isinstance(tracks, list):
for item in tracks:
if not isinstance(item, dict):
continue
direction = _string_value(item.get("direction"))
if not direction:
continue
direction_counts[direction] = direction_counts.get(direction, 0) + 1
total_people = _int_value(payload.get("total_people"))
window_end = _string_value(payload.get("window_end"))
return {
"result_type": PROJECT_TYPE,
"headline": f"Latest window counted {total_people} people",
"last_result_time": window_end,
"metrics": {
"summary_path": str(summary_path) if summary_path else "",
"window_start": _string_value(payload.get("window_start")),
"window_end": window_end,
"total_people": total_people,
"direction_counts": direction_counts,
"age_counts": _map_string_int(payload.get("age_counts")),
"gender_counts": _map_string_int(payload.get("gender_counts")),
"unknown_attributes": _int_value(payload.get("unknown_attributes")),
"recent_window_stats": all_window_stats[:24],
"all_window_stats": all_window_stats,
},
}
def _load_summary_payload(ctx: ManageContext) -> tuple[Path | None, dict | None]:
candidates: list[Path] = []
latest_json = _latest_json_path(ctx)
if latest_json.exists():
candidates.append(latest_json)
window_files = _window_files(ctx)
if window_files:
candidates.extend(window_files[:1])
for candidate in candidates:
try:
payload = json.loads(candidate.read_text(encoding="utf-8"))
except FileNotFoundError:
continue
except json.JSONDecodeError as exc:
raise ValueError(f"invalid summary json: {candidate}") from exc
if isinstance(payload, dict):
return candidate, payload
return None, None
def _load_window_stats(ctx: ManageContext) -> list[dict]:
stats: list[dict] = []
for path in _window_files(ctx):
try:
payload = json.loads(path.read_text(encoding="utf-8"))
except (FileNotFoundError, json.JSONDecodeError):
continue
if not isinstance(payload, dict):
continue
stats.append(
{
"window_start": _string_value(payload.get("window_start")),
"window_end": _string_value(payload.get("window_end")),
"total_people": _int_value(payload.get("total_people")),
"age_counts": _map_string_int(payload.get("age_counts")),
"gender_counts": _map_string_int(payload.get("gender_counts")),
"unknown_attributes": _int_value(payload.get("unknown_attributes")),
}
)
stats.sort(key=lambda item: item["window_end"], reverse=True)
return stats
def _list_result_files(ctx: ManageContext) -> list[dict]:
files: list[dict] = []
for path, label in (
(_latest_json_path(ctx), "Latest Summary"),
(_runtime_log_path(ctx), "Runtime Log"),
):
if path.exists() and path.is_file():
files.append(_build_result_file(ctx, path, label))
for path in _window_files(ctx):
if path.exists() and path.is_file():
files.append(_build_result_file(ctx, path, "Window Summary"))
return files
def _build_result_file(ctx: ManageContext, path: Path, label: str) -> dict:
info = path.stat()
return {
"path": _relative_path(ctx, path),
"name": path.name,
"label": label,
"kind": path.suffix.lstrip(".").lower(),
"size": info.st_size,
"modified_at": datetime.fromtimestamp(info.st_mtime).astimezone().isoformat(),
}
def _output_root(ctx: ManageContext) -> Path:
config = load_config(ctx.config_path)
return resolve_project_path(ctx.project_root, config.runtime.output_dir)
def _rtsp_output_root(ctx: ManageContext) -> Path:
config = load_config(ctx.config_path)
return _output_root(ctx) / config.rtsp.output_subdir
def _latest_json_path(ctx: ManageContext) -> Path:
return _rtsp_output_root(ctx) / "latest.json"
def _windows_dir(ctx: ManageContext) -> Path:
return _rtsp_output_root(ctx) / "windows"
def _runtime_log_path(ctx: ManageContext) -> Path:
return _output_root(ctx) / "rtsp_run.log"
def _window_files(ctx: ManageContext) -> list[Path]:
windows_dir = _windows_dir(ctx)
if not windows_dir.exists():
return []
return sorted(
[path for path in windows_dir.iterdir() if path.is_file()],
key=lambda path: path.name,
reverse=True,
)
def _resolve_sandbox_file(ctx: ManageContext, raw_path: str) -> Path:
relative = raw_path.strip().lstrip("/")
if not relative:
raise ValueError("path is required")
target = (ctx.project_root / relative).resolve()
project_root = ctx.project_root.resolve()
if target != project_root and project_root not in target.parents:
raise ValueError("invalid file path")
if not target.exists() or not target.is_file():
raise FileNotFoundError(relative)
return target
def _relative_path(ctx: ManageContext, target: Path) -> str:
return target.resolve().relative_to(ctx.project_root.resolve()).as_posix()
def _tail_lines(path: Path, line_count: int) -> list[str]:
lines: list[str] = []
with path.open("r", encoding="utf-8") as handle:
for raw_line in handle:
lines.append(raw_line.rstrip("\n"))
if len(lines) > line_count:
lines = lines[1:]
return lines
def _bounded_preview_lines(raw_value: str | None) -> int:
if raw_value is None:
return 200
value = _int_value(raw_value)
if value <= 0:
return 200
return min(value, MAX_PREVIEW_LINES)
def _int_arg(name: str, default: int) -> int:
value = request.args.get(name)
if value is None:
return default
return _int_value(value)
def _string_value(value) -> str:
if value is None:
return ""
return str(value)
def _int_value(value) -> int:
if value is None:
return 0
if isinstance(value, int):
return value
if isinstance(value, float):
return int(value)
try:
return int(str(value).strip())
except ValueError:
return 0
def _map_string_int(value) -> dict[str, int]:
if not isinstance(value, dict):
return {}
return {str(key): _int_value(raw) for key, raw in value.items()}
if __name__ == "__main__":
raise SystemExit(main())

View File

@@ -0,0 +1,105 @@
from __future__ import annotations
from dataclasses import dataclass, field
from pathlib import Path
@dataclass
class YoloConfig:
model_path: str = "yolo11n.pt"
tracker: str = "botsort.yaml"
conf: float = 0.35
iou: float = 0.5
imgsz: int = 1280
device: str = "cuda:0"
@dataclass
class CountingConfig:
line: tuple[float, float, float, float] = (0.1, 0.55, 0.9, 0.55)
line_mode: str = "normalized"
crossing_tolerance: float = 12.0
def to_pixel_line(self, width: int, height: int) -> tuple[float, float, float, float]:
x1, y1, x2, y2 = self.line
if self.line_mode == "pixel":
return x1, y1, x2, y2
return x1 * width, y1 * height, x2 * width, y2 * height
@dataclass
class AttributeConfig:
enabled: bool = True
sample_every_n_frames: int = 12
max_samples_per_track: int = 5
min_person_box_width: int = 80
min_person_box_height: int = 160
person_crop_padding: float = 0.15
detector_backend: str = "retinaface"
enforce_detection: bool = False
@dataclass
class OutputConfig:
save_video: bool = True
save_json: bool = True
save_csv: bool = True
draw_boxes: bool = True
draw_labels: bool = True
@dataclass
class RtspConfig:
sample_interval_seconds: float = 1.0
window_seconds: int = 1800
reconnect_delay_seconds: float = 5.0
stream_open_timeout_seconds: float = 10.0
idle_sleep_seconds: float = 0.05
output_subdir: str = "rtsp_stream"
@dataclass
class RuntimeConfig:
rtsp_url: str = "rtsp://user:password@camera-ip:554/h264/ch1/main/av_stream"
output_dir: str = "outputs"
@dataclass
class AppConfig:
yolo: YoloConfig = field(default_factory=YoloConfig)
counting: CountingConfig = field(default_factory=CountingConfig)
attributes: AttributeConfig = field(default_factory=AttributeConfig)
output: OutputConfig = field(default_factory=OutputConfig)
rtsp: RtspConfig = field(default_factory=RtspConfig)
runtime: RuntimeConfig = field(default_factory=RuntimeConfig)
config_path: Path | None = None
@dataclass
class TrackObservation:
track_id: int
bbox: tuple[int, int, int, int]
confidence: float
center: tuple[float, float]
@dataclass
class CrossingEvent:
track_id: int
direction: str
@dataclass
class AttributeVote:
age: int
age_bucket: str
gender: str
@dataclass
class TrackAttributeSummary:
track_id: int
age: int
age_bucket: str
gender: str
samples_used: int

View File

@@ -0,0 +1,445 @@
from __future__ import annotations
import time
from datetime import datetime, timedelta
from pathlib import Path
from typing import Any
import cv2
from .attributes import AttributeAggregator
from .counting import LineCrossCounter
from .io_utils import (
draw_line,
draw_stats,
draw_tracks,
ensure_dir,
make_video_writer,
write_csv,
write_json,
write_window_json,
)
from .models import AppConfig
from .tracking import extract_person_tracks
SUPPORTED_EXTENSIONS = {".mp4", ".mov", ".mkv", ".avi"}
def discover_videos(root: Path, pattern: str = "*.mp4") -> list[Path]:
if not root.exists():
raise FileNotFoundError(f"Input directory not found: {root}")
videos = [
path
for path in root.rglob(pattern)
if path.is_file() and path.suffix.lower() in SUPPORTED_EXTENSIONS
]
return sorted(videos)
class PeopleFlowPipeline:
def __init__(self, config: AppConfig, output_root: Path) -> None:
self.config = config
self.output_root = ensure_dir(output_root)
self.model = self._load_model()
def _load_model(self) -> Any:
try:
from ultralytics import YOLO
except ImportError as exc:
raise RuntimeError(
"Ultralytics is not installed. Install dependencies with `pip install -r requirements.txt`."
) from exc
return YOLO(self.config.yolo.model_path)
def get_rtsp_output_paths(self) -> dict[str, Path]:
root = ensure_dir(self.output_root / self.config.rtsp.output_subdir)
windows = ensure_dir(root / "windows")
latest_json = root / "latest.json"
return {"root": root, "windows": windows, "latest_json": latest_json}
def process_batch(self, videos: list[Path]) -> dict:
rows: list[dict] = []
for video_path in videos:
rows.append(self.process_video(video_path))
csv_path = self.output_root / "batch_summary.csv"
if self.config.output.save_csv:
csv_rows = [
{
"video_name": row["video_name"],
"video_path": row["video_path"],
"total_people": row["total_people"],
"minor": row["age_counts"]["minor"],
"adult": row["age_counts"]["adult"],
"senior": row["age_counts"]["senior"],
"male": row["gender_counts"]["male"],
"female": row["gender_counts"]["female"],
"unknown_attributes": row["unknown_attributes"],
"json_path": row["json_path"],
"video_output_path": row.get("video_output_path"),
}
for row in rows
]
write_csv(csv_path, csv_rows)
return {"videos": rows, "csv_path": str(csv_path)}
def process_video(self, video_path: Path) -> dict:
if not video_path.exists():
raise FileNotFoundError(f"Video file not found: {video_path}")
capture = cv2.VideoCapture(str(video_path))
if not capture.isOpened():
raise RuntimeError(f"Failed to open video: {video_path}")
width = int(capture.get(cv2.CAP_PROP_FRAME_WIDTH) or 0)
height = int(capture.get(cv2.CAP_PROP_FRAME_HEIGHT) or 0)
fps = float(capture.get(cv2.CAP_PROP_FPS) or 25.0)
pixel_line = self.config.counting.to_pixel_line(width=width, height=height)
video_output_dir = ensure_dir(self.output_root / video_path.stem)
video_output_path = video_output_dir / f"{video_path.stem}.annotated.mp4"
json_path = video_output_dir / f"{video_path.stem}.json"
writer = None
if self.config.output.save_video:
writer = make_video_writer(video_output_path, width=width, height=height, fps=fps)
counter = LineCrossCounter(pixel_line, self.config.counting)
attributes = AttributeAggregator(self.config.attributes)
frame_index = 0
while True:
ok, frame = capture.read()
if not ok:
break
observations = self._track_frame(frame)
for observation in observations:
attributes.maybe_collect(frame=frame, frame_index=frame_index, track=observation)
counter.update(observations)
if writer is not None:
frame_stats = self._build_live_stats(counter, attributes)
annotated = frame.copy()
draw_line(annotated, pixel_line)
if self.config.output.draw_boxes:
draw_tracks(
annotated,
observations=observations,
counted_ids=counter.counted_ids,
draw_labels=self.config.output.draw_labels,
)
draw_stats(annotated, frame_stats)
writer.write(annotated)
frame_index += 1
capture.release()
if writer is not None:
writer.release()
summary = self._finalize_summary(video_path, counter, attributes, json_path)
if not self.config.output.save_video:
summary["video_output_path"] = None
else:
summary["video_output_path"] = str(video_output_path)
return summary
def process_rtsp(self, source: str) -> dict:
rtsp_paths = self.get_rtsp_output_paths()
sample_interval = max(float(self.config.rtsp.sample_interval_seconds), 0.01)
window_seconds = max(int(self.config.rtsp.window_seconds), 1)
reconnect_delay = max(float(self.config.rtsp.reconnect_delay_seconds), 0.1)
open_timeout_seconds = max(float(self.config.rtsp.stream_open_timeout_seconds), 1.0)
idle_sleep = max(float(self.config.rtsp.idle_sleep_seconds), 0.0)
window_index = 0
process_started_at = datetime.now().astimezone()
window_start = datetime.now().astimezone()
window_end = window_start + timedelta(seconds=window_seconds)
last_processed_at = 0.0
last_processed_wall_time: datetime | None = None
next_heartbeat_at = time.monotonic() + 60.0
frame_index = 0
capture = None
pixel_line = None
counter = None
attributes = AttributeAggregator(self.config.attributes)
try:
while True:
now = datetime.now().astimezone()
while now >= window_end:
payload = self._build_rtsp_summary(
source=source,
window_index=window_index,
window_start=window_start,
window_end=window_end,
counter=counter,
attributes=attributes,
)
json_path = write_window_json(
rtsp_paths["windows"],
rtsp_paths["latest_json"],
payload,
window_end,
)
print(f"window_json={json_path}", flush=True)
print(f"window_total_people={payload['total_people']}", flush=True)
window_index += 1
window_start = window_end
window_end = window_start + timedelta(seconds=window_seconds)
if counter is not None:
counter.reset()
attributes.reset()
now = datetime.now().astimezone()
if capture is None or not capture.isOpened():
capture = self._open_rtsp_capture(source, open_timeout_seconds)
if capture is None:
time.sleep(reconnect_delay)
continue
ok, frame = capture.read()
if not ok or frame is None:
capture.release()
capture = None
time.sleep(reconnect_delay)
continue
if pixel_line is None:
height, width = frame.shape[:2]
pixel_line = self.config.counting.to_pixel_line(width=width, height=height)
counter = LineCrossCounter(pixel_line, self.config.counting)
current_time = time.monotonic()
if current_time - last_processed_at < sample_interval:
if idle_sleep > 0:
time.sleep(idle_sleep)
continue
last_processed_at = current_time
observations = self._track_frame(frame)
for observation in observations:
attributes.maybe_collect(frame=frame, frame_index=frame_index, track=observation)
if counter is not None:
counter.update(observations)
if current_time >= next_heartbeat_at:
self._print_rtsp_heartbeat(
process_started_at=process_started_at,
window_index=window_index,
frame_index=frame_index + 1,
counter=counter,
attributes=attributes,
last_processed_wall_time=now,
)
next_heartbeat_at = current_time + 60.0
last_processed_wall_time = now
frame_index += 1
except KeyboardInterrupt:
pass
finally:
if capture is not None:
capture.release()
return {
"rtsp_output_dir": str(rtsp_paths["root"]),
"latest_json": str(rtsp_paths["latest_json"]),
}
def _track_frame(self, frame) -> list:
results = self.model.track(
frame,
persist=True,
tracker=self.config.yolo.tracker,
conf=self.config.yolo.conf,
iou=self.config.yolo.iou,
imgsz=self.config.yolo.imgsz,
device=self.config.yolo.device,
verbose=False,
classes=[0],
)
result = results[0] if isinstance(results, list) else results
return extract_person_tracks(result)
def _open_rtsp_capture(self, source: str, timeout_seconds: float):
capture = cv2.VideoCapture()
open_timeout = getattr(cv2, "CAP_PROP_OPEN_TIMEOUT_MSEC", None)
read_timeout = getattr(cv2, "CAP_PROP_READ_TIMEOUT_MSEC", None)
if open_timeout is not None:
capture.set(open_timeout, timeout_seconds * 1000.0)
if read_timeout is not None:
capture.set(read_timeout, timeout_seconds * 1000.0)
buffersize = getattr(cv2, "CAP_PROP_BUFFERSIZE", None)
if buffersize is not None:
capture.set(buffersize, 1)
capture.open(source)
if capture.isOpened():
return capture
capture.release()
return None
def _build_live_stats(self, counter: LineCrossCounter, attributes: AttributeAggregator) -> dict:
age_counts = {"minor": 0, "adult": 0, "senior": 0}
gender_counts = {"male": 0, "female": 0}
unknown_attributes = 0
for track_id in counter.counted_ids:
summary = attributes.summarize_track(track_id)
if summary is None:
unknown_attributes += 1
continue
age_counts[summary.age_bucket] += 1
gender_counts[summary.gender] += 1
return {
"total_people": counter.total_people,
"age_counts": age_counts,
"gender_counts": gender_counts,
"unknown_attributes": unknown_attributes,
}
def _print_rtsp_heartbeat(
self,
process_started_at: datetime,
window_index: int,
frame_index: int,
counter: LineCrossCounter,
attributes: AttributeAggregator,
last_processed_wall_time: datetime | None,
) -> None:
stats = self._build_live_stats(counter, attributes)
runtime_seconds = int((datetime.now().astimezone() - process_started_at).total_seconds())
last_processed = (
last_processed_wall_time.isoformat(timespec="seconds")
if last_processed_wall_time is not None
else None
)
print(
"heartbeat "
f"runtime_seconds={runtime_seconds} "
f"window_index={window_index} "
f"window_frames={frame_index} "
f"total_people={stats['total_people']} "
f"minor={stats['age_counts']['minor']} "
f"adult={stats['age_counts']['adult']} "
f"senior={stats['age_counts']['senior']} "
f"male={stats['gender_counts']['male']} "
f"female={stats['gender_counts']['female']} "
f"unknown_attributes={stats['unknown_attributes']} "
f"last_processed_at={last_processed}",
flush=True,
)
def _collect_track_summaries(
self,
counter: LineCrossCounter | None,
attributes: AttributeAggregator,
) -> tuple[dict[str, int], dict[str, int], int, list[dict]]:
age_counts = {"minor": 0, "adult": 0, "senior": 0}
gender_counts = {"male": 0, "female": 0}
unknown_attributes = 0
track_summaries: list[dict] = []
if counter is None:
return age_counts, gender_counts, unknown_attributes, track_summaries
for event in counter.crossings:
summary = attributes.summarize_track(event.track_id)
if summary is None:
unknown_attributes += 1
track_summaries.append(
{
"track_id": event.track_id,
"direction": event.direction,
"age": None,
"age_bucket": None,
"gender": None,
"samples_used": 0,
}
)
continue
age_counts[summary.age_bucket] += 1
gender_counts[summary.gender] += 1
track_summaries.append(
{
"track_id": summary.track_id,
"direction": event.direction,
"age": summary.age,
"age_bucket": summary.age_bucket,
"gender": summary.gender,
"samples_used": summary.samples_used,
}
)
return age_counts, gender_counts, unknown_attributes, track_summaries
def _build_rtsp_summary(
self,
source: str,
window_index: int,
window_start: datetime,
window_end: datetime,
counter: LineCrossCounter | None,
attributes: AttributeAggregator,
) -> dict:
age_counts, gender_counts, unknown_attributes, track_summaries = self._collect_track_summaries(
counter,
attributes,
)
total_people = 0 if counter is None else counter.total_people
return {
"source_type": "rtsp",
"source": source,
"window_index": window_index,
"window_start": window_start.isoformat(),
"window_end": window_end.isoformat(),
"window_duration_seconds": int((window_end - window_start).total_seconds()),
"config_path": str(self.config.config_path) if self.config.config_path else None,
"line": {
"coordinates": list(self.config.counting.line),
"mode": self.config.counting.line_mode,
},
"total_people": total_people,
"age_counts": age_counts,
"gender_counts": gender_counts,
"unknown_attributes": unknown_attributes,
"tracks": track_summaries,
}
def _finalize_summary(
self,
video_path: Path,
counter: LineCrossCounter,
attributes: AttributeAggregator,
json_path: Path,
) -> dict:
age_counts, gender_counts, unknown_attributes, track_summaries = self._collect_track_summaries(
counter,
attributes,
)
payload = {
"video_name": video_path.name,
"video_path": str(video_path),
"config_path": str(self.config.config_path) if self.config.config_path else None,
"line": {
"coordinates": list(self.config.counting.line),
"mode": self.config.counting.line_mode,
},
"total_people": counter.total_people,
"age_counts": age_counts,
"gender_counts": gender_counts,
"unknown_attributes": unknown_attributes,
"tracks": track_summaries,
}
if self.config.output.save_json:
write_json(json_path, payload)
payload["json_path"] = str(json_path)
return payload

View File

@@ -0,0 +1,35 @@
from __future__ import annotations
from typing import Any
from .models import TrackObservation
def extract_person_tracks(result: Any) -> list[TrackObservation]:
boxes = getattr(result, "boxes", None)
if boxes is None:
return []
if getattr(boxes, "id", None) is None:
return []
xyxy = boxes.xyxy.int().cpu().tolist()
ids = boxes.id.int().cpu().tolist()
confs = boxes.conf.cpu().tolist()
classes = boxes.cls.int().cpu().tolist()
observations: list[TrackObservation] = []
for bbox, track_id, confidence, class_id in zip(xyxy, ids, confs, classes, strict=False):
if int(class_id) != 0:
continue
x1, y1, x2, y2 = bbox
center_x = (x1 + x2) / 2.0
center_y = (y1 + y2) / 2.0
observations.append(
TrackObservation(
track_id=int(track_id),
bbox=(int(x1), int(y1), int(x2), int(y2)),
confidence=float(confidence),
center=(center_x, center_y),
)
)
return observations

View File

@@ -0,0 +1,164 @@
from __future__ import annotations
import json
from pathlib import Path
import yaml
from src.people_flow.manage_api import create_app
def build_client(project_root: Path):
config_path = project_root / "config" / "local.yaml"
config_path.parent.mkdir(parents=True, exist_ok=True)
config_path.write_text(
"runtime:\n"
" rtsp_url: rtsp://before-update\n"
" output_dir: outputs\n"
"rtsp:\n"
" output_subdir: rtsp_stream\n",
encoding="utf-8",
)
rtsp_dir = project_root / "outputs" / "rtsp_stream"
windows_dir = rtsp_dir / "windows"
windows_dir.mkdir(parents=True, exist_ok=True)
latest_payload = {
"source_type": "rtsp",
"window_start": "2026-04-16T09:30:00+08:00",
"window_end": "2026-04-16T10:00:00+08:00",
"total_people": 7,
"age_counts": {"minor": 1, "adult": 5, "senior": 1},
"gender_counts": {"male": 4, "female": 3},
"unknown_attributes": 2,
"tracks": [
{"track_id": 1, "direction": "in"},
{"track_id": 2, "direction": "out"},
{"track_id": 3, "direction": "in"},
],
}
(rtsp_dir / "latest.json").write_text(
json.dumps(latest_payload),
encoding="utf-8",
)
(windows_dir / "stats_2026-04-16_09-00-00.json").write_text(
json.dumps(
{
"window_start": "2026-04-16T09:00:00+08:00",
"window_end": "2026-04-16T09:30:00+08:00",
"total_people": 5,
"age_counts": {"minor": 0, "adult": 4, "senior": 1},
"gender_counts": {"male": 2, "female": 3},
"unknown_attributes": 1,
}
),
encoding="utf-8",
)
(windows_dir / "stats_2026-04-16_09-30-00.json").write_text(
json.dumps(latest_payload),
encoding="utf-8",
)
(project_root / "outputs" / "rtsp_run.log").write_text("rtsp ok\n", encoding="utf-8")
app = create_app(config_path)
app.testing = True
return app.test_client(), config_path
def test_get_manage_health(tmp_path: Path):
client, _ = build_client(tmp_path)
response = client.get("/api/manage/health")
assert response.status_code == 200
assert response.json["status"] == "ok"
assert response.json["project_type"] == "people_flow_project"
assert response.json["runtime_status"] == "running"
def test_get_manage_config(tmp_path: Path):
client, config_path = build_client(tmp_path)
response = client.get("/api/manage/config")
assert response.status_code == 200
assert response.json["config_path"] == str(config_path)
assert response.json["runtime"]["rtsp_url"] == "rtsp://before-update"
assert response.json["rtsp"]["output_subdir"] == "rtsp_stream"
def test_put_manage_config_updates_rtsp_url(tmp_path: Path):
client, config_path = build_client(tmp_path)
response = client.put(
"/api/manage/config",
json={"rtsp_url": "rtsp://after-update"},
)
assert response.status_code == 200
assert response.json["runtime"]["rtsp_url"] == "rtsp://after-update"
saved = yaml.safe_load(config_path.read_text(encoding="utf-8"))
assert saved["runtime"]["rtsp_url"] == "rtsp://after-update"
def test_get_manage_summary(tmp_path: Path):
client, _ = build_client(tmp_path)
response = client.get("/api/manage/summary")
assert response.status_code == 200
assert response.json["result_type"] == "people_flow_project"
assert response.json["last_result_time"] == "2026-04-16T10:00:00+08:00"
assert response.json["metrics"]["total_people"] == 7
assert response.json["metrics"]["direction_counts"] == {"in": 2, "out": 1}
assert response.json["metrics"]["recent_window_stats"][0]["window_end"] == "2026-04-16T10:00:00+08:00"
def test_get_manage_windows(tmp_path: Path):
client, _ = build_client(tmp_path)
response = client.get("/api/manage/windows?page=1&page_size=1")
assert response.status_code == 200
assert response.json["total"] == 2
assert response.json["page"] == 1
assert response.json["page_size"] == 1
assert response.json["items"][0]["window_end"] == "2026-04-16T10:00:00+08:00"
assert response.json["items"][0]["total_people"] == 7
def test_get_manage_files(tmp_path: Path):
client, _ = build_client(tmp_path)
response = client.get("/api/manage/files")
assert response.status_code == 200
assert {item["path"] for item in response.json["files"]} == {
"outputs/rtsp_run.log",
"outputs/rtsp_stream/latest.json",
"outputs/rtsp_stream/windows/stats_2026-04-16_09-00-00.json",
"outputs/rtsp_stream/windows/stats_2026-04-16_09-30-00.json",
}
def test_get_manage_files_preview(tmp_path: Path):
client, _ = build_client(tmp_path)
response = client.get(
"/api/manage/files/preview?path=outputs/rtsp_stream/latest.json&lines=1"
)
assert response.status_code == 200
assert response.json["path"] == "outputs/rtsp_stream/latest.json"
assert response.json["count"] == 1
assert "total_people" in response.json["lines"][0]
def test_get_manage_files_download(tmp_path: Path):
client, _ = build_client(tmp_path)
response = client.get("/api/manage/files/download?path=outputs/rtsp_run.log")
assert response.status_code == 200
assert response.data == b"rtsp ok\n"

View File

@@ -0,0 +1 @@

View File

@@ -0,0 +1 @@