fix: PaddleOCR model downloads failing silently in CI

Remove continue-on-error from model download steps so failures are
visible. Add retry logic with exponential backoff (3 attempts), increase
curl timeouts (30s connect, 600s max), bump cache key to v2, and fail
the verify step when no models are found.

Also remove unused cache-binding-artifact action.

Author: Goldziher

.github/actions/cache-binding-artifact/action.yml

@@ -41,7 +41,7 @@ runs:
       id: cache-models
       with:
         path: ~/.cache/kreuzberg/paddle-ocr
-        key: ${{ inputs.cache-key-suffix }}-${{ runner.os }}-${{ runner.arch }}-v1
+        key: ${{ inputs.cache-key-suffix }}-${{ runner.os }}-${{ runner.arch }}-v2
         restore-keys: |
           ${{ inputs.cache-key-suffix }}-${{ runner.os }}-${{ runner.arch }}-
           ${{ inputs.cache-key-suffix }}-${{ runner.os }}-
@@ -50,7 +50,6 @@ runs:
     - name: Download detection model (det)
       if: contains(inputs.models, 'det') && steps.cache-models.outputs.cache-hit != 'true'
       shell: bash
-      continue-on-error: true
       run: |
         MODEL_URL="https://huggingface.co/Kreuzberg/paddleocr-onnx-models/resolve/main/ch_PP-OCRv4_det_infer.onnx"
         CACHE_DIR="$HOME/.cache/kreuzberg/paddle-ocr"
@@ -60,18 +59,27 @@ runs:
         echo "Downloading detection model from $MODEL_URL"
         mkdir -p "$MODEL_DIR"
 
-        if curl -f -L --progress-bar --connect-timeout 10 --max-time 300 \
-          -o "$MODEL_FILE" "$MODEL_URL"; then
-          echo "Detection model downloaded successfully ($(du -h "$MODEL_FILE" | cut -f1))"
-        else
-          echo "Warning: Failed to download detection model, CI will continue without it"
-          rm -f "$MODEL_FILE"
-        fi
+        for attempt in 1 2 3; do
+          if [ $attempt -gt 1 ]; then
+            backoff=$((5 * 3 ** (attempt - 2)))
+            echo "Retry attempt $attempt/$3 after ${backoff}s backoff..."
+            sleep $backoff
+          fi
+
+          if curl -f -L --progress-bar --connect-timeout 30 --max-time 600 \
+            -o "$MODEL_FILE" "$MODEL_URL"; then
+            echo "Detection model downloaded successfully ($(du -h "$MODEL_FILE" | cut -f1))"
+            exit 0
+          fi
+        done
+
+        echo "ERROR: Failed to download detection model after 3 attempts"
+        rm -f "$MODEL_FILE"
+        exit 1
 
     - name: Download classification model (cls)
       if: contains(inputs.models, 'cls') && steps.cache-models.outputs.cache-hit != 'true'
       shell: bash
-      continue-on-error: true
       run: |
         MODEL_URL="https://huggingface.co/Kreuzberg/paddleocr-onnx-models/resolve/main/ch_ppocr_mobile_v2.0_cls_infer.onnx"
         CACHE_DIR="$HOME/.cache/kreuzberg/paddle-ocr"
@@ -81,18 +89,27 @@ runs:
         echo "Downloading classification model from $MODEL_URL"
         mkdir -p "$MODEL_DIR"
 
-        if curl -f -L --progress-bar --connect-timeout 10 --max-time 300 \
-          -o "$MODEL_FILE" "$MODEL_URL"; then
-          echo "Classification model downloaded successfully ($(du -h "$MODEL_FILE" | cut -f1))"
-        else
-          echo "Warning: Failed to download classification model, CI will continue without it"
-          rm -f "$MODEL_FILE"
-        fi
+        for attempt in 1 2 3; do
+          if [ $attempt -gt 1 ]; then
+            backoff=$((5 * 3 ** (attempt - 2)))
+            echo "Retry attempt $attempt/$3 after ${backoff}s backoff..."
+            sleep $backoff
+          fi
+
+          if curl -f -L --progress-bar --connect-timeout 30 --max-time 600 \
+            -o "$MODEL_FILE" "$MODEL_URL"; then
+            echo "Classification model downloaded successfully ($(du -h "$MODEL_FILE" | cut -f1))"
+            exit 0
+          fi
+        done
+
+        echo "ERROR: Failed to download classification model after 3 attempts"
+        rm -f "$MODEL_FILE"
+        exit 1
 
     - name: Download recognition model (rec)
       if: contains(inputs.models, 'rec') && steps.cache-models.outputs.cache-hit != 'true'
       shell: bash
-      continue-on-error: true
       run: |
         MODEL_URL="https://huggingface.co/Kreuzberg/paddleocr-onnx-models/resolve/main/en_PP-OCRv4_rec_infer.onnx"
         CACHE_DIR="$HOME/.cache/kreuzberg/paddle-ocr"
@@ -102,13 +119,23 @@ runs:
         echo "Downloading recognition model from $MODEL_URL"
         mkdir -p "$MODEL_DIR"
 
-        if curl -f -L --progress-bar --connect-timeout 10 --max-time 300 \
-          -o "$MODEL_FILE" "$MODEL_URL"; then
-          echo "Recognition model downloaded successfully ($(du -h "$MODEL_FILE" | cut -f1))"
-        else
-          echo "Warning: Failed to download recognition model, CI will continue without it"
-          rm -f "$MODEL_FILE"
-        fi
+        for attempt in 1 2 3; do
+          if [ $attempt -gt 1 ]; then
+            backoff=$((5 * 3 ** (attempt - 2)))
+            echo "Retry attempt $attempt/$3 after ${backoff}s backoff..."
+            sleep $backoff
+          fi
+
+          if curl -f -L --progress-bar --connect-timeout 30 --max-time 600 \
+            -o "$MODEL_FILE" "$MODEL_URL"; then
+            echo "Recognition model downloaded successfully ($(du -h "$MODEL_FILE" | cut -f1))"
+            exit 0
+          fi
+        done
+
+        echo "ERROR: Failed to download recognition model after 3 attempts"
+        rm -f "$MODEL_FILE"
+        exit 1
 
     - name: Verify downloaded models
       id: verify-models
@@ -142,9 +169,9 @@ runs:
         fi
 
         if [ ${#AVAILABLE_MODELS[@]} -eq 0 ]; then
-          echo "⚠ No models found in cache directory"
+          echo "ERROR: No models found in cache directory after download"
           echo "available-models=" >> $GITHUB_OUTPUT
-          exit 0
+          exit 1
         fi
 
         AVAILABLE_MODELS_STR=$(IFS=, ; echo "${AVAILABLE_MODELS[*]}")