Spaces:
Runtime error
Runtime error
Commit ·
b50c572
1
Parent(s): 85c59d5
Simplify Docker setup and dependencies for minimal API version
Browse files- Dockerfile +2 -33
- entrypoint.sh +3 -108
- requirements.txt +1 -9
Dockerfile
CHANGED
|
@@ -3,14 +3,7 @@ FROM python:3.10-slim
|
|
| 3 |
# Install system dependencies
|
| 4 |
RUN apt-get update && \
|
| 5 |
apt-get install -y --no-install-recommends \
|
| 6 |
-
build-essential \
|
| 7 |
curl \
|
| 8 |
-
libgl1-mesa-glx \
|
| 9 |
-
libglib2.0-0 \
|
| 10 |
-
wget \
|
| 11 |
-
git \
|
| 12 |
-
libopenblas-dev \
|
| 13 |
-
libgomp1 \
|
| 14 |
&& rm -rf /var/lib/apt/lists/*
|
| 15 |
|
| 16 |
# Set working directory
|
|
@@ -20,20 +13,11 @@ WORKDIR /app
|
|
| 20 |
ENV PYTHONUNBUFFERED=1 \
|
| 21 |
PYTHONDONTWRITEBYTECODE=1 \
|
| 22 |
UPLOAD_FOLDER=/tmp/pdf_uploads \
|
| 23 |
-
OUTPUT_FOLDER=/tmp/pdf_output
|
| 24 |
-
HF_HUB_DISABLE_PROGRESS_BARS=1 \
|
| 25 |
-
HF_HUB_ENABLE_HF_TRANSFER=0
|
| 26 |
|
| 27 |
# Create necessary directories and set permissions
|
| 28 |
RUN mkdir -p /tmp/pdf_uploads /tmp/pdf_output \
|
| 29 |
-
&&
|
| 30 |
-
&& mkdir -p /tmp/models/MFR/unimernet \
|
| 31 |
-
&& mkdir -p /tmp/models/table/rapid \
|
| 32 |
-
&& mkdir -p /tmp/models/layout/doclayout \
|
| 33 |
-
&& chmod -R 777 /tmp/pdf_uploads /tmp/pdf_output /tmp/models
|
| 34 |
-
|
| 35 |
-
# Create magic-pdf config directories
|
| 36 |
-
RUN mkdir -p /root/.config/magic_pdf
|
| 37 |
|
| 38 |
# Copy requirements and app files
|
| 39 |
COPY requirements.txt /app/
|
|
@@ -42,23 +26,8 @@ COPY entrypoint.sh /app/
|
|
| 42 |
RUN chmod +x /app/entrypoint.sh
|
| 43 |
|
| 44 |
# Install Python dependencies
|
| 45 |
-
RUN pip install --no-cache-dir --upgrade pip setuptools wheel
|
| 46 |
RUN pip install --no-cache-dir -r requirements.txt
|
| 47 |
|
| 48 |
-
# Copy default magic-pdf config
|
| 49 |
-
RUN echo '{"device-mode":"cpu","layout-config":{"model":"doclayout_yolo","enable":true},"formula-config":{"mfd_model":"yolo_v8_mfd","mfr_model":"unimernet_small","enable":false},"table-config":{"model":"rapid_table","sub_model":"slanet_plus","enable":false}}' > /root/.config/magic_pdf/magic-pdf.json
|
| 50 |
-
|
| 51 |
-
# Download model files directly
|
| 52 |
-
RUN curl -L https://huggingface.co/marcosremar2/mineru-models/resolve/main/doclayout_yolo.pt -o /tmp/models/layout/doclayout/doclayout_yolo.pt
|
| 53 |
-
RUN curl -L https://huggingface.co/marcosremar2/mineru-models/resolve/main/yolo_v8_ft.pt -o /tmp/models/MFD/YOLO/yolo_v8_ft.pt
|
| 54 |
-
|
| 55 |
-
# Verify model files were downloaded correctly
|
| 56 |
-
RUN ls -la /tmp/models/layout/doclayout/doclayout_yolo.pt
|
| 57 |
-
RUN ls -la /tmp/models/MFD/YOLO/yolo_v8_ft.pt
|
| 58 |
-
|
| 59 |
-
# Check if magic-pdf exists but don't require it to work yet
|
| 60 |
-
RUN pip show minerupdf || echo "MinerU PDF will be installed at runtime"
|
| 61 |
-
|
| 62 |
# Expose port
|
| 63 |
EXPOSE 7860
|
| 64 |
|
|
|
|
| 3 |
# Install system dependencies
|
| 4 |
RUN apt-get update && \
|
| 5 |
apt-get install -y --no-install-recommends \
|
|
|
|
| 6 |
curl \
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7 |
&& rm -rf /var/lib/apt/lists/*
|
| 8 |
|
| 9 |
# Set working directory
|
|
|
|
| 13 |
ENV PYTHONUNBUFFERED=1 \
|
| 14 |
PYTHONDONTWRITEBYTECODE=1 \
|
| 15 |
UPLOAD_FOLDER=/tmp/pdf_uploads \
|
| 16 |
+
OUTPUT_FOLDER=/tmp/pdf_output
|
|
|
|
|
|
|
| 17 |
|
| 18 |
# Create necessary directories and set permissions
|
| 19 |
RUN mkdir -p /tmp/pdf_uploads /tmp/pdf_output \
|
| 20 |
+
&& chmod -R 777 /tmp/pdf_uploads /tmp/pdf_output
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
|
| 22 |
# Copy requirements and app files
|
| 23 |
COPY requirements.txt /app/
|
|
|
|
| 26 |
RUN chmod +x /app/entrypoint.sh
|
| 27 |
|
| 28 |
# Install Python dependencies
|
|
|
|
| 29 |
RUN pip install --no-cache-dir -r requirements.txt
|
| 30 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31 |
# Expose port
|
| 32 |
EXPOSE 7860
|
| 33 |
|
entrypoint.sh
CHANGED
|
@@ -5,115 +5,10 @@ set -e
|
|
| 5 |
echo "Starting entrypoint.sh..."
|
| 6 |
echo "User: $(whoami)"
|
| 7 |
echo "Current directory: $(pwd)"
|
| 8 |
-
echo "Directory listing:"
|
| 9 |
-
ls -la
|
| 10 |
|
| 11 |
-
#
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
# Verify directories existence and permissions
|
| 16 |
-
echo "Checking directories..."
|
| 17 |
-
for dir in /tmp/pdf_uploads /tmp/pdf_output /tmp/models/MFD/YOLO /tmp/models/MFR/unimernet /tmp/models/table/rapid /tmp/models/layout/doclayout
|
| 18 |
-
do
|
| 19 |
-
if [ ! -d "$dir" ]; then
|
| 20 |
-
echo "Creating directory: $dir"
|
| 21 |
-
mkdir -p "$dir"
|
| 22 |
-
fi
|
| 23 |
-
chmod -R 777 "$dir"
|
| 24 |
-
echo "Directory $dir is ready"
|
| 25 |
-
done
|
| 26 |
-
|
| 27 |
-
# Copy config file to all possible locations to ensure it's found
|
| 28 |
-
echo "Setting up magic-pdf config..."
|
| 29 |
-
CONFIG_CONTENT='{
|
| 30 |
-
"device-mode": "cpu",
|
| 31 |
-
"layout-config": {
|
| 32 |
-
"model": "doclayout_yolo",
|
| 33 |
-
"enable": true
|
| 34 |
-
},
|
| 35 |
-
"formula-config": {
|
| 36 |
-
"mfd_model": "yolo_v8_mfd",
|
| 37 |
-
"mfr_model": "unimernet_small",
|
| 38 |
-
"enable": false
|
| 39 |
-
},
|
| 40 |
-
"table-config": {
|
| 41 |
-
"model": "rapid_table",
|
| 42 |
-
"sub_model": "slanet_plus",
|
| 43 |
-
"enable": false
|
| 44 |
-
}
|
| 45 |
-
}'
|
| 46 |
-
|
| 47 |
-
# Create config in all possible locations
|
| 48 |
-
mkdir -p ~/.config/magic_pdf
|
| 49 |
-
echo "$CONFIG_CONTENT" > ~/.config/magic_pdf/magic-pdf.json
|
| 50 |
-
echo "$CONFIG_CONTENT" > ~/magic-pdf.json
|
| 51 |
-
echo "$CONFIG_CONTENT" > /app/magic-pdf.json
|
| 52 |
-
echo "$CONFIG_CONTENT" > /home/user/magic-pdf.json
|
| 53 |
-
echo "$CONFIG_CONTENT" > /root/.config/magic_pdf/magic-pdf.json
|
| 54 |
-
|
| 55 |
-
# Download model function with validation
|
| 56 |
-
function download_model_with_validation() {
|
| 57 |
-
local model_path=$1
|
| 58 |
-
local model_url=$2
|
| 59 |
-
local max_attempts=3
|
| 60 |
-
local attempt=1
|
| 61 |
-
|
| 62 |
-
echo "Checking model file: $model_path"
|
| 63 |
-
|
| 64 |
-
if [ ! -f "$model_path" ] || [ ! -s "$model_path" ]; then
|
| 65 |
-
while [ $attempt -le $max_attempts ]; do
|
| 66 |
-
echo "Downloading model attempt $attempt/$max_attempts: $model_url"
|
| 67 |
-
curl -L --retry 5 --retry-delay 2 "$model_url" -o "$model_path"
|
| 68 |
-
|
| 69 |
-
# Verify file exists and has content
|
| 70 |
-
if [ -f "$model_path" ] && [ -s "$model_path" ]; then
|
| 71 |
-
size=$(du -h "$model_path" | cut -f1)
|
| 72 |
-
echo "✅ Model downloaded successfully ($size): $model_path"
|
| 73 |
-
return 0
|
| 74 |
-
else
|
| 75 |
-
echo "❌ Download failed or file is empty. Retrying..."
|
| 76 |
-
rm -f "$model_path" 2>/dev/null
|
| 77 |
-
attempt=$((attempt+1))
|
| 78 |
-
sleep 2
|
| 79 |
-
fi
|
| 80 |
-
done
|
| 81 |
-
|
| 82 |
-
echo "❌ Failed to download model after $max_attempts attempts: $model_url"
|
| 83 |
-
exit 1
|
| 84 |
-
else
|
| 85 |
-
size=$(du -h "$model_path" | cut -f1)
|
| 86 |
-
echo "✅ Model already exists ($size): $model_path"
|
| 87 |
-
fi
|
| 88 |
-
}
|
| 89 |
-
|
| 90 |
-
# Download and verify all required models
|
| 91 |
-
echo "Verifying model files..."
|
| 92 |
-
download_model_with_validation "/tmp/models/layout/doclayout/doclayout_yolo.pt" "https://huggingface.co/marcosremar2/mineru-models/resolve/main/doclayout_yolo.pt"
|
| 93 |
-
download_model_with_validation "/tmp/models/MFD/YOLO/yolo_v8_ft.pt" "https://huggingface.co/marcosremar2/mineru-models/resolve/main/yolo_v8_ft.pt"
|
| 94 |
-
|
| 95 |
-
# List all model files for verification
|
| 96 |
-
echo "Model files verification:"
|
| 97 |
-
find /tmp/models -type f -exec ls -la {} \;
|
| 98 |
-
|
| 99 |
-
# Verify magic-pdf exists and is executable
|
| 100 |
-
echo "Checking magic-pdf installation..."
|
| 101 |
-
if command -v magic-pdf &> /dev/null; then
|
| 102 |
-
echo "magic-pdf found. Testing version:"
|
| 103 |
-
magic-pdf --version || echo "Warning: magic-pdf command exists but might not be functioning properly."
|
| 104 |
-
else
|
| 105 |
-
echo "magic-pdf command not found. Installing..."
|
| 106 |
-
pip install --no-cache-dir minerupdf==1.3.6
|
| 107 |
-
if command -v magic-pdf &> /dev/null; then
|
| 108 |
-
echo "magic-pdf installed successfully. Testing version:"
|
| 109 |
-
magic-pdf --version || echo "Warning: magic-pdf command was installed but might not be functioning properly."
|
| 110 |
-
else
|
| 111 |
-
echo "Warning: Failed to install magic-pdf."
|
| 112 |
-
echo "PATH: $PATH"
|
| 113 |
-
pip list | grep miner
|
| 114 |
-
# Continue anyway
|
| 115 |
-
fi
|
| 116 |
-
fi
|
| 117 |
|
| 118 |
# Start the Flask application
|
| 119 |
echo "Starting Flask application..."
|
|
|
|
| 5 |
echo "Starting entrypoint.sh..."
|
| 6 |
echo "User: $(whoami)"
|
| 7 |
echo "Current directory: $(pwd)"
|
|
|
|
|
|
|
| 8 |
|
| 9 |
+
# Create necessary directories
|
| 10 |
+
mkdir -p /tmp/pdf_uploads /tmp/pdf_output
|
| 11 |
+
chmod -R 777 /tmp/pdf_uploads /tmp/pdf_output
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
|
| 13 |
# Start the Flask application
|
| 14 |
echo "Starting Flask application..."
|
requirements.txt
CHANGED
|
@@ -1,12 +1,4 @@
|
|
| 1 |
flask==2.3.3
|
| 2 |
flask-cors==4.0.0
|
| 3 |
werkzeug==2.3.7
|
| 4 |
-
|
| 5 |
-
numpy>=1.20.0
|
| 6 |
-
requests>=2.25.0
|
| 7 |
-
opencv-python-headless>=4.5.0
|
| 8 |
-
torch>=1.8.0
|
| 9 |
-
torchvision>=0.9.0
|
| 10 |
-
transformers>=4.15.0
|
| 11 |
-
huggingface_hub>=0.11.0
|
| 12 |
-
minerupdf==1.3.6
|
|
|
|
| 1 |
flask==2.3.3
|
| 2 |
flask-cors==4.0.0
|
| 3 |
werkzeug==2.3.7
|
| 4 |
+
requests>=2.25.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|