Hugging Face's logo

Spaces:
point9
/
Invoice_Digitization_Agent
Sleeping

App Files Community
dipan004 commited on
Commit
d3bf819
·
verified ·
1 Parent(s): 02a4759

Update Dockerfile

Browse files
Files changed (1) hide show
  1. Dockerfile +42 -35
Dockerfile CHANGED
@@ -1,10 +1,13 @@
 
 
 
1
  FROM python:3.10-slim
2
 
3
  WORKDIR /app
4
 
5
- # --------------------------------------------
6
- # Install system dependencies + OCR libraries
7
- # --------------------------------------------
8
  RUN apt-get update && apt-get install -y --no-install-recommends \
9
  sqlite3 \
10
  ca-certificates \
@@ -17,53 +20,55 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
17
  libxrender-dev \
18
  libgl1 \
19
  build-essential \
 
 
 
20
  && rm -rf /var/lib/apt/lists/*
21
 
22
  # Verify Tesseract installation
23
  RUN tesseract --version
24
 
25
- # --------------------------------------------
26
- # Copy requirements
27
- # --------------------------------------------
28
  COPY requirements.txt .
29
 
30
- # --------------------------------------------
31
- # Upgrade pip + build tools
32
- # --------------------------------------------
33
  RUN pip install --no-cache-dir --upgrade pip setuptools wheel cython
34
 
35
- # --------------------------------------------
36
- # 🔧 Fix NumPy/Pandas binary compatibility
37
- # 1️⃣ Install NumPy first (stable ABI)
38
- # 2️⃣ Install Pandas from source against installed NumPy
39
- # --------------------------------------------
40
- RUN pip install --no-cache-dir --upgrade --force-reinstall numpy==1.24.4
41
- RUN pip install --no-cache-dir --upgrade --force-reinstall --no-binary :all: pandas==2.0.3
42
 
43
- # --------------------------------------------
44
- # Install remaining dependencies WITHOUT touching numpy/pandas
45
- # --------------------------------------------
46
- RUN pip install --no-cache-dir --upgrade --ignore-installed -r requirements.txt
47
 
48
- # Upgrade Gemini SDK for v1 API
49
- RUN pip install --no-cache-dir --upgrade google-generativeai google-ai-generativelanguage
 
 
50
 
51
- # Install EasyOCR + dependencies
 
 
 
52
  RUN pip install --no-cache-dir easyocr opencv-python-headless Pillow pytesseract
53
 
54
- # --------------------------------------------
55
  # Copy application code
56
- # --------------------------------------------
57
  COPY . .
58
 
59
- # --------------------------------------------
60
  # Create necessary directories
61
- # --------------------------------------------
62
  RUN mkdir -p /app/data/logs /app/data/docs /app/backend/app/db && chmod -R 777 /app/data
63
 
64
- # --------------------------------------------
65
  # Create __init__.py files
66
- # --------------------------------------------
67
  RUN touch backend/__init__.py \
68
  && touch backend/feature_builder/__init__.py \
69
  && touch backend/app/__init__.py \
@@ -73,26 +78,28 @@ RUN touch backend/__init__.py \
73
  && touch backend/app/db/__init__.py \
74
  && touch backend/ingest/__init__.py
75
 
 
76
  # Verify agent files exist
 
77
  RUN test -f backend/app/agent/agent_orchestrator.py || \
78
  (echo "ERROR: agent_orchestrator.py not found! Add it before building." && exit 1)
79
 
80
- # --------------------------------------------
81
  # Initialize database
82
- # --------------------------------------------
83
  COPY backend/app/db/db_init.py backend/app/db/db_init.py
84
  RUN echo "🗄️ Initializing database during build..." && \
85
  python backend/app/db/db_init.py && \
86
  echo "✅ Database initialized successfully!"
87
 
88
- # --------------------------------------------
89
  # Expose port
90
- # --------------------------------------------
91
  EXPOSE 7860
92
 
93
- # --------------------------------------------
94
  # Startup script
95
- # --------------------------------------------
96
  RUN echo '#!/bin/bash\n\
97
  echo "🔍 Checking database..."\n\
98
  python backend/app/db/db_init.py\n\
 
1
+ # ===============================
2
+ # Base image
3
+ # ===============================
4
  FROM python:3.10-slim
5
 
6
  WORKDIR /app
7
 
8
+ # ===============================
9
+ # System dependencies + OCR libraries
10
+ # ===============================
11
  RUN apt-get update && apt-get install -y --no-install-recommends \
12
  sqlite3 \
13
  ca-certificates \
 
20
  libxrender-dev \
21
  libgl1 \
22
  build-essential \
23
+ gcc \
24
+ g++ \
25
+ python3-dev \
26
  && rm -rf /var/lib/apt/lists/*
27
 
28
  # Verify Tesseract installation
29
  RUN tesseract --version
30
 
31
+ # ===============================
32
+ # Copy requirements and upgrade pip
33
+ # ===============================
34
  COPY requirements.txt .
35
 
 
 
 
36
  RUN pip install --no-cache-dir --upgrade pip setuptools wheel cython
37
 
38
+ # ===============================
39
+ # Install NumPy first (prebuilt wheel)
40
+ # ===============================
41
+ RUN pip install --no-cache-dir numpy==1.24.4
 
 
 
42
 
43
+ # ===============================
44
+ # Install Pandas from prebuilt wheel (avoid --no-binary)
45
+ # ===============================
46
+ RUN pip install --no-cache-dir pandas==2.0.3 --only-binary=:all:
47
 
48
+ # ===============================
49
+ # Install remaining dependencies
50
+ # ===============================
51
+ RUN pip install --no-cache-dir -r requirements.txt
52
 
53
+ # ===============================
54
+ # Optional: Gemini SDK, EasyOCR
55
+ # ===============================
56
+ RUN pip install --no-cache-dir --upgrade google-generativeai google-ai-generativelanguage
57
  RUN pip install --no-cache-dir easyocr opencv-python-headless Pillow pytesseract
58
 
59
+ # ===============================
60
  # Copy application code
61
+ # ===============================
62
  COPY . .
63
 
64
+ # ===============================
65
  # Create necessary directories
66
+ # ===============================
67
  RUN mkdir -p /app/data/logs /app/data/docs /app/backend/app/db && chmod -R 777 /app/data
68
 
69
+ # ===============================
70
  # Create __init__.py files
71
+ # ===============================
72
  RUN touch backend/__init__.py \
73
  && touch backend/feature_builder/__init__.py \
74
  && touch backend/app/__init__.py \
 
78
  && touch backend/app/db/__init__.py \
79
  && touch backend/ingest/__init__.py
80
 
81
+ # ===============================
82
  # Verify agent files exist
83
+ # ===============================
84
  RUN test -f backend/app/agent/agent_orchestrator.py || \
85
  (echo "ERROR: agent_orchestrator.py not found! Add it before building." && exit 1)
86
 
87
+ # ===============================
88
  # Initialize database
89
+ # ===============================
90
  COPY backend/app/db/db_init.py backend/app/db/db_init.py
91
  RUN echo "🗄️ Initializing database during build..." && \
92
  python backend/app/db/db_init.py && \
93
  echo "✅ Database initialized successfully!"
94
 
95
+ # ===============================
96
  # Expose port
97
+ # ===============================
98
  EXPOSE 7860
99
 
100
+ # ===============================
101
  # Startup script
102
+ # ===============================
103
  RUN echo '#!/bin/bash\n\
104
  echo "🔍 Checking database..."\n\
105
  python backend/app/db/db_init.py\n\