Merge pull request #1 from chris-rutkowski/v102

V102

Author: chris-rutkowski

README.md

@@ -14,7 +14,26 @@
 
 ## 🛠️ Usage
 
-### 1. **Create an Ignore File**
+### 1. **Add the GitHub Action**
+Create a GitHub Actions workflow in `.github/workflows/duplicate_guard.yml`:
+
+```yaml
+name: Duplicate Guard
+on:
+  pull_request:
+    branches:
+      - main
+  workflow_dispatch:
+
+jobs:
+  duplicate_guard:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Duplicate Guard
+        uses: chris-rutkowski/duplicate-guard@v1.0.2
+```
+
+### 2. **Create an ignore file** (optional)
 Add a `duplicate_guard.ignore` file to the root of your repository to define patterns for files or directories to exclude from duplicate checks. The syntax follows `.gitignore` conventions.
 
 **Example `duplicate_guard.ignore`:**
@@ -26,38 +45,31 @@ logs/*
 
 ---
 
-### 2. **Add the GitHub Action**
-Create a GitHub Actions workflow in `.github/workflows/duplicate_guard.yml`:
+## ♻️ Find existing duplicates
+
+Run the action manually using the `workflow_dispatch` event to scan and find duplicates in your repository.
 
 ```yaml
 name: Duplicate Guard
 on:
-  pull_request:
-    branches:
-      - master
   workflow_dispatch:
+  pull_request:
 
-jobs:
-  filesize_guard:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Duplicate Guard
-        uses: chris-rutkowski/duplicate-guard@v1.0.1
+...
 ```
 
 ---
 
 ## ⚙️ Configuration
 
-### **Specify a Custom Ignore File Path**
-If your `duplicate_guard.ignore` file is not in the root directory, specify its location using the `ignore_file` input:
+### **Specify a custom ignore file path**
 
 ```yaml
 steps:
   - name: Duplicate Guard
-    uses: chris-rutkowski/filesize-guard@v1.0.1
+    uses: chris-rutkowski/duplicate-guard@v1.0.2
       with:
-        ignore_file: ./my/path/my_filesize_guard.ignore
+        ignore_file: ./my/path/my_duplicate_guard.ignore
 ```
 
 ---

action.yml

@@ -14,14 +14,22 @@ runs:
       uses: actions/checkout@v4
 
     - name: Get changed files
+      if: ${{ github.event_name == 'pull_request' }}
       id: changed-files
       uses: tj-actions/changed-files@v45
       with:
         json: true
         write_output_files: true
         safe_output: false
 
+    - name: Get all files
+      if: ${{ github.event_name != 'pull_request' }}
+      run: |
+        mkdir -p .github/outputs/
+        find . -type f | sed 's|^\./||' | jq -R . | jq -s . > .github/outputs/all_repo_files.json
+      shell: bash
+
     - name: Run Duplicate Guard
       run: |
-        python3 ${GITHUB_ACTION_PATH}/duplicate_guard.py ${{ inputs.ignore_file }} .github/outputs/modified_files.json .github/outputs/added_files.json
+        python3 ${GITHUB_ACTION_PATH}/duplicate_guard.py ${{ inputs.ignore_file }} .github/outputs/modified_files.json .github/outputs/added_files.json .github/outputs/all_repo_files.json
       shell: bash

duplicate_guard.py

@@ -4,9 +4,21 @@
 import os
 import sys
 
+DEFAULT_IGNORE_FILE = "./duplicate_guard.ignore"
+
 def load_ignore_patterns(ignore_file):
+    default_patterns = [".git/*"]
+
+    if not os.path.exists(ignore_file):
+        if ignore_file == DEFAULT_IGNORE_FILE:
+            return default_patterns
+
+        print(f"Error: The specified ignore file '{ignore_file}' does not exist.", file=sys.stderr)
+        sys.exit(1)
+
     with open(ignore_file, "r") as f:
-        return [line.strip() for line in f if line.strip() and not line.startswith("#")]
+        patterns = [line.strip() for line in f if line.strip() and not line.startswith("#")]
+    return default_patterns + patterns
 
 def should_ignore(file, patterns):
     return any(fnmatch.fnmatch(file, pattern) for pattern in patterns)
@@ -31,6 +43,8 @@ def get_all_repository_files(ignore_patterns):
 def load_files_from_json(file_paths):
     files = []
     for file_path in file_paths:
+        if not os.path.exists(file_path):
+            continue
         with open(file_path, "r") as f:
             files.extend(json.load(f))
     return files
@@ -43,7 +57,8 @@ def load_files_from_json(file_paths):
 checksums = {}
 for file in get_all_repository_files(ignore_patterns):
     checksum = calculate_checksum(file)
-    checksums[checksum] = file
+    if checksum not in checksums:
+        checksums[checksum] = file
 print(f"Done, {len(checksums)} checksums")
 
 # Step 2: Check new/modified files against the repository and themselves
@@ -54,11 +69,8 @@ def load_files_from_json(file_paths):
         continue
 
     if should_ignore(file, ignore_patterns):
-        print(f"Ignoring: '{file}'")
         continue
 
-    print(f"Processing: '{file}'")
-
     checksum = calculate_checksum(file)
 
     if checksum in checksums: