Update Dockerfile

Dockerfile

@@ -1,62 +1,100 @@
-# ---- Base image ----
-FROM python:3.11-slim
+# We use Python 3.10 because it has the best pre-built wheels for Llama
+FROM python:3.10-slim
 
-# Workdir inside the container
 WORKDIR /app
 
-# ---- System dependencies ----
+# 1. Install minimal tools (Git/Curl) but NO heavy compilers
 RUN apt-get update && apt-get install -y \
-    build-essential \
     curl \
     git \
-    unzip \
- && rm -rf /var/lib/apt/lists/*
+    build-essential \
+    && rm -rf /var/lib/apt/lists/*
 
+# 2. Upgrade pip (Critical for finding the right wheels)
+RUN pip install --upgrade pip
 
-# ---- Python deps ----
-COPY requirements.txt ./
+# 3. Install LLM Engine (Pre-built)
+# This installs the tool on the hard drive. It takes space, but ZERO CPU to run.
+# The --prefer-binary flag ensures we don't try to compile it.
+RUN pip install llama-cpp-python \
+    --no-cache-dir \
+    --prefer-binary \
+    --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cpu
 
-# Torch / sentence-transformers like having the CPU wheel index explicitly
-RUN pip install --no-cache-dir \
-    --extra-index-url https://download.pytorch.org/whl/cpu \
-    -r requirements.txt
+# 4. Install other requirements
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
 
-# ---- NLTK data (punkt + stopwords) ----
-RUN mkdir -p /usr/local/share/nltk_data
+# 5. Download NLTK data
+RUN python -c "import nltk; nltk.download('punkt'); nltk.download('punkt_tab'); nltk.download('stopwords')"
 
-# punkt tokenizer
-# RUN curl -L "https://raw.githubusercontent.com/nltk/nltk_data/gh-pages/packages/tokenizers/punkt.zip" \
-#       -o /tmp/punkt.zip && \
-#     unzip /tmp/punkt.zip -d /usr/local/share/nltk_data && \
-#     rm /tmp/punkt.zip
-# punkt_tab tokenizer (for NLTK >= 3.9)
-RUN curl -L "https://raw.githubusercontent.com/nltk/nltk_data/gh-pages/packages/tokenizers/punkt_tab.zip" \
-      -o /tmp/punkt_tab.zip && \
-    unzip /tmp/punkt_tab.zip -d /usr/local/share/nltk_data && \
-    rm /tmp/punkt_tab.zip
+COPY . .
 
+# 6. Run the Unified App (Back to app.py!)
+CMD ["streamlit", "run", "app2.0.py", "--server.port=7860", "--server.address=0.0.0.0"]
 
-# stopwords corpus
-RUN curl -L "https://raw.githubusercontent.com/nltk/nltk_data/gh-pages/packages/corpora/stopwords.zip" \
-      -o /tmp/stopwords.zip && \
-    unzip /tmp/stopwords.zip -d /usr/local/share/nltk_data && \
-    rm /tmp/stopwords.zip
 
-# ---- Copy app code ----
-# If you only want app.py + data, you can narrow this, but copying all is fine.
-COPY . .
 
-# ---- Hugging Face port wiring ----
-ENV PORT=7860
-EXPOSE 7860
 
-# Optional healthcheck; HF will just ignore failures but nice to have
-HEALTHCHECK CMD curl --fail http://localhost:${PORT}/_stcore/health || exit 1
+# # ---- Base image ----
+# FROM python:3.11-slim
+
+# # Workdir inside the container
+# WORKDIR /app
+
+# # ---- System dependencies ----
+# RUN apt-get update && apt-get install -y \
+#     build-essential \
+#     curl \
+#     git \
+#     unzip \
+#  && rm -rf /var/lib/apt/lists/*
+
+
+# # ---- Python deps ----
+# COPY requirements.txt ./
+
+# # Torch / sentence-transformers like having the CPU wheel index explicitly
+# RUN pip install --no-cache-dir \
+#     --extra-index-url https://download.pytorch.org/whl/cpu \
+#     -r requirements.txt
+
+# # ---- NLTK data (punkt + stopwords) ----
+# RUN mkdir -p /usr/local/share/nltk_data
+
+# # punkt tokenizer
+# # RUN curl -L "https://raw.githubusercontent.com/nltk/nltk_data/gh-pages/packages/tokenizers/punkt.zip" \
+# #       -o /tmp/punkt.zip && \
+# #     unzip /tmp/punkt.zip -d /usr/local/share/nltk_data && \
+# #     rm /tmp/punkt.zip
+# # punkt_tab tokenizer (for NLTK >= 3.9)
+# RUN curl -L "https://raw.githubusercontent.com/nltk/nltk_data/gh-pages/packages/tokenizers/punkt_tab.zip" \
+#       -o /tmp/punkt_tab.zip && \
+#     unzip /tmp/punkt_tab.zip -d /usr/local/share/nltk_data && \
+#     rm /tmp/punkt_tab.zip
+
+
+# # stopwords corpus
+# RUN curl -L "https://raw.githubusercontent.com/nltk/nltk_data/gh-pages/packages/corpora/stopwords.zip" \
+#       -o /tmp/stopwords.zip && \
+#     unzip /tmp/stopwords.zip -d /usr/local/share/nltk_data && \
+#     rm /tmp/stopwords.zip
+
+# # ---- Copy app code ----
+# # If you only want app.py + data, you can narrow this, but copying all is fine.
+# COPY . .
+
+# # ---- Hugging Face port wiring ----
+# ENV PORT=7860
+# EXPOSE 7860
+
+# # Optional healthcheck; HF will just ignore failures but nice to have
+# HEALTHCHECK CMD curl --fail http://localhost:${PORT}/_stcore/health || exit 1
 
-# ---- Run Streamlit ----
-ENTRYPOINT ["bash", "-c", "streamlit run app.py \
-    --server.port=${PORT} \
-    --server.address=0.0.0.0 \
-    --server.enableCORS=false \
-    --server.enableXsrfProtection=false"]
+# # ---- Run Streamlit ----
+# ENTRYPOINT ["bash", "-c", "streamlit run app.py \
+#     --server.port=${PORT} \
+#     --server.address=0.0.0.0 \
+#     --server.enableCORS=false \
+#     --server.enableXsrfProtection=false"]