test draft

Author: chienyuanchang

python/content_understanding_client.py

@@ -323,7 +323,7 @@ def begin_analyze(self, analyzer_id: str, file_location: str) -> Response:
                     ]
                 }
                 headers = {"Content-Type": "application/json"}
-            elif file_path.is_file() and self.is_supported_type_by_file_path(file_path):
+            elif file_path.is_file():
                 with open(file_location, "rb") as file:
                     data = file.read()
                 headers = {"Content-Type": "application/octet-stream"}

tools/README.md

@@ -0,0 +1,60 @@
+# test_notebooks.py
+
+This script is designed to automatically execute Jupyter notebooks in the `notebooks/` directory (or a specified directory) and report their execution status. It is useful for validating that all notebooks in the repository run successfully from start to finish, which is especially helpful for CI/CD pipelines or for contributors to verify their changes.
+
+## Features
+- **Automatic Discovery:** Recursively scans a directory for `.ipynb` files (excluding hidden files).
+- **Selective Skipping:** Supports a skip list to exclude specific notebooks from execution (e.g., those requiring manual input).
+- **Execution Reporting:** Prints a summary of successful and failed notebooks, including error messages for failures.
+- **Command Line Usage:** Can run all notebooks in a directory or a specified list of notebook files.
+
+## Usage
+
+### Run All Notebooks in a Directory
+
+```bash
+python3 tools/test_notebooks.py
+```
+This will scan the `notebooks/` directory by default, skipping any notebooks listed in the `skip_list` variable.
+
+### Run Specific Notebooks
+
+```bash
+python3 tools/test_notebooks.py notebooks/example1.ipynb notebooks/example2.ipynb
+```
+This will execute only the specified notebooks.
+
+## Skip List
+You can modify the `skip_list` variable in the script to add or remove notebooks that should be skipped during execution. The skip list can contain full paths or substrings.
+
+## Dependencies
+- Python 3
+- `nbformat`
+- `nbconvert`
+
+Install dependencies with:
+```bash
+pip3 install nbformat nbconvert
+```
+
+## Exit Codes
+- Returns `0` if all notebooks succeed.
+- Returns `1` if any notebook fails or if no notebooks are found.
+
+## Example Output
+```
+🔍 Scanning for notebooks in: /path/to/notebooks
+▶️ Running: /path/to/notebooks/example.ipynb
+✅ Success: /path/to/notebooks/example.ipynb
+...
+🧾 Notebook Execution Summary
+✅ 3 succeeded
+❌ 1 failed
+🚨 Failed notebooks:
+ - /path/to/notebooks/failing.ipynb
+   ↳ Traceback (most recent call last): ...
+```
+
+## Notes
+- Notebooks that require manual input or special setup should be added to the skip list.
+- The script is intended for use in development, CI, or pre-merge validation workflows.

tools/test_notebooks.py

@@ -0,0 +1,121 @@
+import os
+import sys
+from typing import Tuple, Optional, List
+
+import nbformat
+from nbconvert.preprocessors import ExecutePreprocessor
+
+
+SINGLE_NOTEBOOK_TIMEOUT = 1200
+
+
+def should_skip(notebook_path: str, skip_list: List[str]) -> bool:
+    return any(skip in notebook_path for skip in skip_list)
+
+
+def run_notebook(notebook_path: str, root: str) -> Tuple[bool, Optional[str]]:
+    """Execute a single notebook."""
+    try:
+        with open(notebook_path, encoding="utf-8") as f:
+            nb = nbformat.read(f, as_version=4)
+
+        ep = ExecutePreprocessor(
+            timeout=SINGLE_NOTEBOOK_TIMEOUT,
+            kernel_name="python3")
+        ep.preprocess(nb, {"metadata": {"path": root}})
+        return True, None
+    except Exception as e:
+        return False, str(e)
+
+
+def run_all_notebooks(path: str = ".", skip_list: List[str] = None) -> None:
+    abs_path = os.path.abspath(path)
+    print(f"🔍 Scanning for notebooks in: {abs_path}\n")
+
+    skip_list = skip_list or []
+
+    notebook_found: int = 0
+    success_notebooks: List[str] = []
+    failed_notebooks: List[Tuple[str, str]] = []
+
+    for root, _, files in os.walk(abs_path):
+        for file in files:
+            if file.endswith(".ipynb") and not file.startswith("."):
+                notebook_path = os.path.join(root, file)
+
+                if should_skip(notebook_path, skip_list):
+                    print(f"⏭️ Skipped: {notebook_path}")
+                    continue
+
+                notebook_found += 1
+                print(f"▶️ Running: {notebook_path}")
+                success, error = run_notebook(notebook_path, root)
+
+                if success:
+                    print(f"✅ Success: {notebook_path}\n")
+                    success_notebooks.append(notebook_path)
+                else:
+                    print(f"❌ Failed: {notebook_path}\nError: {error}\n")
+                    failed_notebooks.append((notebook_path, error))
+
+    # 📋 Summary
+    print("🧾 Notebook Execution Summary")
+    print(f"✅ {len(success_notebooks)} succeeded")
+    print(f"❌ {len(failed_notebooks)} failed\n")
+
+    if failed_notebooks:
+        print("🚨 Failed notebooks:")
+        for nb, error in failed_notebooks:
+            last_line = error.strip().splitlines()[-1] if error else "Unknown error"
+            print(f" - {nb}\n   ↳ {last_line}")
+        sys.exit(1)
+
+    if notebook_found == 0:
+        print("❌ No notebooks were found. Check the folder path or repo contents.")
+        sys.exit(1)
+
+    print("🏁 All notebooks completed successfully.")
+
+
+if __name__ == "__main__":
+    args: List[str] = sys.argv[1:]
+
+    # NOTE: Define skip list (can use full paths or substrings)
+    skip_list = [
+        "build_person_directory.ipynb",  # Skip due to "new_face_image_path" needed to be added manually
+    ]
+
+    if not args:
+        run_all_notebooks("notebooks", skip_list=skip_list)
+    else:
+        failed: List[Tuple[str, str]] = []
+        for notebook_path in args:
+            if should_skip(notebook_path, skip_list):
+                print(f"⏭️ Skipped: {notebook_path}")
+                continue
+
+            if notebook_path.endswith(".ipynb") and os.path.isfile(notebook_path):
+                print(f"▶️ Running: {notebook_path}")
+                success, error = run_notebook(notebook_path, os.path.dirname(notebook_path))
+                if success:
+                    print(f"✅ Success: {notebook_path}\n")
+                else:
+                    print(f"❌ Failed: {notebook_path}\nError: {error}\n")
+                    failed.append((notebook_path, error))
+            else:
+                print(f"⚠️ Not a valid notebook file: {notebook_path}")
+                failed.append((notebook_path, "Invalid path or not a .ipynb file"))
+
+        # Summary
+        print("🧾 Execution Summary")
+        print(f"✅ {len(args) - len(failed)} succeeded")
+        print(f"❌ {len(failed)} failed")
+
+        if failed:
+            print("🚨 Failed notebooks:")
+            for nb, error in failed:
+                last_line = error.strip().splitlines()[-1] if error else "Unknown error"
+                print(f" - {nb}\n   ↳ {last_line}")
+            sys.exit(1)
+        else:
+            print("🏁 All selected notebooks completed successfully.")