From fd240e2533aeb44e4ea83693989f106dc638105a Mon Sep 17 00:00:00 2001
From: matt <matt.adam.wisnowski@gmail.com>
Date: Fri, 17 Oct 2025 17:29:17 -0700
Subject: [PATCH] feat: bake similarity cache into Docker images

- Copy card_files/ to /.defaults/card_files/ during build
- Entrypoint checks for baked-in cache before downloading
- Fallback to GitHub download if not present in image
- Reduces startup time and bandwidth usage
---
 Dockerfile    |  6 ++++++
 entrypoint.sh | 16 ++++++++++++----
 2 files changed, 18 insertions(+), 4 deletions(-)

diff --git a/Dockerfile b/Dockerfile
index 06b2381..7f6f0ce 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -32,6 +32,12 @@ COPY config/ ./config/
 COPY config/ /.defaults/config/
 RUN mkdir -p owned_cards
 
+# Copy similarity cache if available (pre-built during CI)
+# Store in /.defaults/card_files so it persists after volume mount  
+RUN mkdir -p /.defaults/card_files
+# Copy entire card_files directory (will include cache if present, empty if not)
+COPY card_files/ /.defaults/card_files/
+
 # Create necessary directories as mount points
 RUN mkdir -p deck_files logs csv_files card_files config /.defaults
 
diff --git a/entrypoint.sh b/entrypoint.sh
index a59e62f..7529ce0 100644
--- a/entrypoint.sh
+++ b/entrypoint.sh
@@ -6,11 +6,19 @@ seed_defaults() {
     # Ensure base config and data directories exist
     mkdir -p /app/config /app/config/card_lists /app/config/themes /app/card_files
 
-    # Download pre-built similarity cache from GitHub if not present
+    # Copy pre-built similarity cache from image defaults if available, otherwise download
     if [ ! -f /app/card_files/similarity_cache.parquet ]; then
-        echo "Downloading similarity cache from GitHub..."
-        wget -q https://raw.githubusercontent.com/mwisnowski/mtg_python_deckbuilder/similarity-cache-data/card_files/similarity_cache.parquet -O /app/card_files/similarity_cache.parquet 2>/dev/null || echo "Warning: Could not download similarity cache"
-        wget -q https://raw.githubusercontent.com/mwisnowski/mtg_python_deckbuilder/similarity-cache-data/card_files/similarity_cache_metadata.json -O /app/card_files/similarity_cache_metadata.json 2>/dev/null || true
+        # First try to copy from baked-in defaults (included in Docker image during CI builds)
+        if [ -f /.defaults/card_files/similarity_cache.parquet ]; then
+            echo "Copying pre-built similarity cache from image..."
+            cp /.defaults/card_files/similarity_cache.parquet /app/card_files/ 2>/dev/null || true
+            cp /.defaults/card_files/similarity_cache_metadata.json /app/card_files/ 2>/dev/null || true
+        else
+            # Fallback: download from GitHub similarity-cache-data branch
+            echo "Downloading similarity cache from GitHub..."
+            wget -q https://raw.githubusercontent.com/mwisnowski/mtg_python_deckbuilder/similarity-cache-data/card_files/similarity_cache.parquet -O /app/card_files/similarity_cache.parquet 2>/dev/null || echo "Warning: Could not download similarity cache"
+            wget -q https://raw.githubusercontent.com/mwisnowski/mtg_python_deckbuilder/similarity-cache-data/card_files/similarity_cache_metadata.json -O /app/card_files/similarity_cache_metadata.json 2>/dev/null || true
+        fi
     fi
 
     # Copy from baked-in defaults if targets are missing