src/python-doctr/.github/workflows/references.yml

name: references

on:
  push:
    branches: main
  pull_request:
    branches: main

jobs:
  train-char-classification:
    runs-on: ${{ matrix.os }}
    strategy:
      fail-fast: false
      matrix:
        os: [ubuntu-latest]
        python: ["3.10"]
    steps:
      - uses: actions/checkout@v5
      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: ${{ matrix.python }}
          architecture: x64

      - name: Cache python modules
        uses: actions/cache@v4
        with:
          path: ~/.cache/pip
          key: ${{ runner.os }}-pkg-deps-${{ matrix.python }}-${{ hashFiles('requirements-pt.txt') }}-${{ hashFiles('references/requirements.txt') }}
          restore-keys: |
            ${{ runner.os }}-pkg-deps-${{ matrix.python }}-${{ hashFiles('requirements-pt.txt') }}-
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          pip install -e .[viz,html] --upgrade
          pip install -r references/requirements.txt
          sudo apt-get update && sudo apt-get install fonts-freefont-ttf -y
      - name: Train for a short epoch
        run: python references/classification/train_character.py vit_s -b 32 --val-samples 1 --train-samples 1 --epochs 1

  train-orientation-classification:
    runs-on: ${{ matrix.os }}
    strategy:
      fail-fast: false
      matrix:
        os: [ubuntu-latest]
        python: ["3.10"]
    steps:
      - uses: actions/checkout@v5
      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: ${{ matrix.python }}
          architecture: x64
      - name: Cache python modules
        uses: actions/cache@v4
        with:
          path: ~/.cache/pip
          key: ${{ runner.os }}-pkg-deps-${{ matrix.python }}-${{ hashFiles('requirements-pt.txt') }}-${{ hashFiles('references/requirements.txt') }}
          restore-keys: |
            ${{ runner.os }}-pkg-deps-${{ matrix.python }}-${{ hashFiles('requirements-pt.txt') }}-
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          pip install -e .[viz,html] --upgrade
          pip install -r references/requirements.txt
      - name: Download and extract detection toy set
        run: |
          wget https://github.com/mindee/doctr/releases/download/v0.3.1/toy_detection_set-bbbb4243.zip
          sudo apt-get update && sudo apt-get install unzip -y
          unzip toy_detection_set-bbbb4243.zip -d det_set
      - name: Download and extract recognition toy set
        run: |
          wget https://github.com/mindee/doctr/releases/download/v0.3.1/toy_recogition_set-036a4d80.zip
          sudo apt-get update && sudo apt-get install unzip -y
          unzip toy_recogition_set-036a4d80.zip -d reco_set
      - name: Train for a short epoch (document orientation)
        run: python references/classification/train_orientation.py resnet18 --type page --train_path ./det_set --val_path ./det_set  -b 2 --epochs 1
      - name: Train for a short epoch (crop orientation)
        run: python references/classification/train_orientation.py resnet18 --type crop --train_path ./reco_set --val_path ./reco_set  -b 4 --epochs 1

  train-text-recognition:
    runs-on: ${{ matrix.os }}
    strategy:
      fail-fast: false
      matrix:
        os: [ubuntu-latest]
        python: ["3.10"]
    steps:
      - uses: actions/checkout@v5
      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: ${{ matrix.python }}
          architecture: x64
      - name: Cache python modules
        uses: actions/cache@v4
        with:
          path: ~/.cache/pip
          key: ${{ runner.os }}-pkg-deps-${{ matrix.python }}-${{ hashFiles('requirements-pt.txt') }}-${{ hashFiles('references/requirements.txt') }}
          restore-keys: |
            ${{ runner.os }}-pkg-deps-${{ matrix.python }}-${{ hashFiles('requirements-pt.txt') }}-
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          pip install -e .[viz,html] --upgrade
          pip install -r references/requirements.txt
      - name: Download and extract toy set
        run: |
          wget https://github.com/mindee/doctr/releases/download/v0.3.1/toy_recogition_set-036a4d80.zip
          sudo apt-get update && sudo apt-get install unzip -y
          unzip toy_recogition_set-036a4d80.zip -d reco_set
      - name: Train for a short epoch
        run: python references/recognition/train.py crnn_mobilenet_v3_small --train_path ./reco_set --val_path ./reco_set -b 4 --epochs 1

  evaluate-text-recognition:
    runs-on: ${{ matrix.os }}
    strategy:
      fail-fast: false
      matrix:
        os: [ubuntu-latest]
        python: ["3.10"]
    steps:
      - uses: actions/checkout@v5
      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: ${{ matrix.python }}
          architecture: x64
      - name: Cache python modules
        uses: actions/cache@v4
        with:
          path: ~/.cache/pip
          key: ${{ runner.os }}-pkg-deps-${{ matrix.python }}-${{ hashFiles('requirements-pt.txt') }}
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          pip install -e .[viz,html] --upgrade
      - name: Evaluate text recognition
        run: python references/recognition/evaluate.py crnn_mobilenet_v3_small --dataset SVT -b 32

  latency-text-recognition:
    runs-on: ${{ matrix.os }}
    strategy:
      fail-fast: false
      matrix:
        os: [ubuntu-latest]
        python: ["3.10"]
    steps:
      - uses: actions/checkout@v5
      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: ${{ matrix.python }}
          architecture: x64
      - name: Cache python modules
        uses: actions/cache@v4
        with:
          path: ~/.cache/pip
          key: ${{ runner.os }}-pkg-deps-${{ matrix.python }}-${{ hashFiles('requirements-pt.txt') }}
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          pip install -e .[viz,html] --upgrade
      - name: Benchmark latency
        run: python references/recognition/latency.py crnn_mobilenet_v3_small --it 5

  train-text-detection:
    runs-on: ${{ matrix.os }}
    strategy:
      fail-fast: false
      matrix:
        os: [ubuntu-latest]
        python: ["3.10"]
    steps:
      - uses: actions/checkout@v5
      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: ${{ matrix.python }}
          architecture: x64
      - name: Cache python modules
        uses: actions/cache@v4
        with:
          path: ~/.cache/pip
          key: ${{ runner.os }}-pkg-deps-${{ matrix.python }}-${{ hashFiles('requirements-pt.txt') }}-${{ hashFiles('references/requirements.txt') }}
          restore-keys: |
            ${{ runner.os }}-pkg-deps-${{ matrix.python }}-${{ hashFiles('requirements-pt.txt') }}-
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          pip install -e .[viz,html] --upgrade
          pip install -r references/requirements.txt
      - name: Download and extract toy set
        run: |
          wget https://github.com/mindee/doctr/releases/download/v0.3.1/toy_detection_set-bbbb4243.zip
          sudo apt-get update && sudo apt-get install unzip -y
          unzip toy_detection_set-bbbb4243.zip -d det_set
      - name: Train for a short epoch
        run: python references/detection/train.py db_mobilenet_v3_large --train_path ./det_set --val_path ./det_set -b 2 --epochs 1

  evaluate-text-detection:
    runs-on: ${{ matrix.os }}
    strategy:
      fail-fast: false
      matrix:
        os: [ubuntu-latest]
        python: ["3.10"]
    steps:
      - uses: actions/checkout@v5
      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: ${{ matrix.python }}
          architecture: x64
      - name: Cache python modules
        uses: actions/cache@v4
        with:
          path: ~/.cache/pip
          key: ${{ runner.os }}-pkg-deps-${{ matrix.python }}-${{ hashFiles('requirements-pt.txt') }}
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          pip install -e .[viz,html] --upgrade
          pip install -r references/requirements.txt
      - name: Evaluate text detection
        run: python references/detection/evaluate.py db_mobilenet_v3_large

  latency-text-detection:
    runs-on: ${{ matrix.os }}
    strategy:
      fail-fast: false
      matrix:
        os: [ubuntu-latest]
        python: ["3.10"]
    steps:
      - uses: actions/checkout@v5
      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: ${{ matrix.python }}
          architecture: x64
      - name: Cache python modules
        uses: actions/cache@v4
        with:
          path: ~/.cache/pip
          key: ${{ runner.os }}-pkg-deps-${{ matrix.python }}-${{ hashFiles('requirements-pt.txt') }}
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          pip install -e .[viz,html] --upgrade
      - name: Benchmark latency
        run: python references/detection/latency.py db_mobilenet_v3_large --it 5 --size 512