Merge pull request #7 from pelikhan/copilot/fix-a5e52bc0-33dc-4256-9143-f7d56da9e69a

Add video URL parameter support with optional instructions and dedicated workflow

Author: pelikhan

.github/workflows/ci.yml

@@ -6,11 +6,6 @@ on:
   pull_request:
     branches:
       - main
-  workflow_dispatch:
-    inputs:
-      github_issue:
-        description: "GitHub issue number to comment on"
-        required: true
 permissions:
   contents: read
   issues: write
@@ -55,6 +50,6 @@ jobs:
             genaiscript-
       - uses: ./
         with:
-          github_issue: 1
+          github_issue: "1"
           github_token: ${{ secrets.GITHUB_TOKEN }}
-          debug: "script"
+          debug: "script"

.github/workflows/genai-video-issue-analyzer.yml

@@ -1,7 +1,18 @@
-name: genai video issue analyzer
+name: Video Analyzer
 on:
   issues:
     types: [opened, edited]
+  workflow_dispatch:
+    inputs:
+      video_url:
+        description: 'Direct video URL to analyze'
+        required: true
+        type: string
+      instructions:
+        description: 'Custom prompting instructions for the video'
+        required: false
+        default: 'Analyze the video and provide a summary of its content. Extract list of followup subissues if any. The transcript is your primary source of text information, ignore text in images.'
+        type: string
 permissions:
     contents: read
     issues: write
@@ -31,4 +42,6 @@ jobs:
       - uses: pelikhan/action-genai-video-issue-analyzer@main
         with:
           github_issue: ${{ github.event.issue.number }}
-          github_token: ${{ secrets.GITHUB_TOKEN }}
+          github_token: ${{ secrets.GITHUB_TOKEN }}
+          video_url: ${{ github.event.inputs.video_url }}
+          instructions: ${{ github.event.inputs.instructions }}

README.md

@@ -1,7 +1,9 @@
 # GitHub Action Video Issue Analyzer
 
-This GitHub Action runs all video assets in an issue body through a LLM model to analyze the content.
-The default behavior is to summarize and extract task items but this can be customized through the `prompt` input.
+This GitHub Action runs all video assets in an issue body through a LLM model to analyze the content, or can analyze a direct video URL when triggered via workflow_dispatch.
+The default behavior is to summarize and extract task items but this can be customized through the `instructions` input.
+
+The action outputs the analysis results to the GitHub Step Summary for easy viewing in the Actions tab.
 
 ## Inputs
 
@@ -22,6 +24,7 @@ The default behavior is to summarize and extract task items but this can be cust
 | `azure_ai_inference_api_version` | Azure Serverless OpenAI API version | false |  |
 | `azure_ai_inference_api_credentials` | Azure Serverless OpenAI API credentials type | false |  |
 | `github_token` | GitHub token with `models: read` permission at least (https://microsoft.github.io/genaiscript/reference/github-actions/#github-models-permissions). | false |  |
+| `video_url` | Direct video URL to analyze (alternative to extracting from issue body). Used when triggered via workflow_dispatch. | false |  |
 
 ## Outputs
 
@@ -30,6 +33,8 @@ The default behavior is to summarize and extract task items but this can be cust
 
 | `text` | The generated text output. |
 
+**Note**: The action also outputs the analysis results to the GitHub Step Summary (`$GITHUB_STEP_SUMMARY`) for easy viewing in the Actions tab.
+
 ## Usage
 
 Add the following to your step in your workflow file.
@@ -62,6 +67,8 @@ It will launch a whisper service in a container that can be used by genaiscript.
 
 Save the following in `.github/workflows/genai-video-issue-analyzer.yml` file:
 
+### For Issue-based Analysis (automatic trigger)
+
 ```yaml
 name: genai video issue analyzer
 on:
@@ -104,6 +111,114 @@ jobs:
           github_token: ${{ secrets.GITHUB_TOKEN }}
 ```
 
+### For Direct Video URL Analysis (manual trigger)
+
+```yaml
+name: genai video issue analyzer
+on:
+  workflow_dispatch:
+    inputs:
+      video_url:
+        description: 'Direct video URL to analyze'
+        required: true
+        type: string
+      instructions:
+        description: 'Custom prompting instructions for the video'
+        required: false
+        default: 'Analyze the video and provide a summary of its content. Extract list of followup subissues if any. The transcript is your primary source of text information, ignore text in images.'
+        type: string
+permissions:
+    contents: read
+    models: read
+jobs:
+  genai-video-analyze:
+    runs-on: ubuntu-latest
+    services:
+      whisper:
+        image: onerahmet/openai-whisper-asr-webservice:latest
+        env:
+          ASR_MODEL: base
+          ASR_ENGINE: openai_whisper
+        ports:
+          - 9000:9000
+        options: >-
+          --health-cmd "curl -f http://localhost:9000/docs || exit 1"
+          --health-interval 10s
+          --health-timeout 5s
+          --health-retries 5
+          --health-start-period 20s
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/cache@v4
+        with:
+          path: .genaiscript/cache/**
+          key: genaiscript-${{ github.run_id }}
+          restore-keys: |
+            genaiscript-
+      - uses: pelikhan/action-genai-video-issue-analyzer@v0 # update to the major version you want to use
+        with:
+          github_token: ${{ secrets.GITHUB_TOKEN }}
+          video_url: ${{ github.event.inputs.video_url }}
+          instructions: ${{ github.event.inputs.instructions }}
+```
+
+### Combined Workflow (both triggers)
+
+```yaml
+name: genai video issue analyzer
+on:
+  issues:
+    types: [opened, edited]
+  workflow_dispatch:
+    inputs:
+      video_url:
+        description: 'Direct video URL to analyze'
+        required: true
+        type: string
+      instructions:
+        description: 'Custom prompting instructions for the video'
+        required: false
+        default: 'Analyze the video and provide a summary of its content. Extract list of followup subissues if any. The transcript is your primary source of text information, ignore text in images.'
+        type: string
+permissions:
+    contents: read
+    issues: write
+    models: read
+concurrency:
+    group: ${{ github.workflow }}-${{ github.ref }}
+    cancel-in-progress: true
+jobs:
+  genai-video-analyze:
+    runs-on: ubuntu-latest
+    services:
+      whisper:
+        image: onerahmet/openai-whisper-asr-webservice:latest
+        env:
+          ASR_MODEL: base
+          ASR_ENGINE: openai_whisper
+        ports:
+          - 9000:9000
+        options: >-
+          --health-cmd "curl -f http://localhost:9000/docs || exit 1"
+          --health-interval 10s
+          --health-timeout 5s
+          --health-retries 5
+          --health-start-period 20s
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/cache@v4
+        with:
+          path: .genaiscript/cache/**
+          key: genaiscript-${{ github.run_id }}
+          restore-keys: |
+            genaiscript-
+      - uses: pelikhan/action-genai-video-issue-analyzer@v0 # update to the major version you want to use
+        with:
+          github_token: ${{ secrets.GITHUB_TOKEN }}
+          video_url: ${{ github.event.inputs.video_url }}
+          instructions: ${{ github.event.inputs.instructions }}
+```
+
 ## Development
 
 This action was automatically generated by [GenAIScript](https://microsoft.github.io/genaiscript/reference/github-actions) from the script metadata.

action.yml

@@ -57,6 +57,13 @@ inputs:
       (https://microsoft.github.io/genaiscript/reference/github-actions/#github\
       -models-permissions)."
     required: false
+  github_issue:
+    description: "GitHub issue number to analyze. If not provided, the action will\
+      \ analyze the current pull request."
+    required: false
+  video_url:
+    description: "Direct video URL to analyze (alternative to extracting from issue body)"
+    required: false
 outputs:
   text:
     description: The generated text output.

genaisrc/action-video-issue-analyzer.genai.mts

@@ -8,34 +8,46 @@ script({
       default:
         "Analyze the video and provide a summary of its content. Extract list of followup subissues if any. The transcript is your primary source of text information, ignore text in images.",
     },
+    videoUrl: {
+      type: "string",
+      description: "Direct video URL to analyze (alternative to extracting from issue body)",
+    },
   },
 });
 
 const { dbg, output, vars } = env;
-const issue = await github.getIssue();
-if (!issue)
-  throw new Error(
-    "No issue found in the context. This action requires an issue to be present.",
-  );
-const { instructions } = vars as { instructions: string };
-if (!instructions)
-  throw new Error(
-    "No instructions provided. Please provide instructions to process the video.",
-  );
+const { instructions, videoUrl } = vars as { instructions?: string; videoUrl?: string };
+
+// Use default instructions if not provided
+const finalInstructions = instructions || 
+  "Analyze the video and provide a summary of its content. Extract list of followup subissues if any. The transcript is your primary source of text information, ignore text in images.";
+
+// Process direct video URL if provided
+if (videoUrl) {
+  dbg(`Processing direct video URL: ${videoUrl}`);
+  await processDirectVideoUrl(videoUrl);
+} else {
+  // Fallback to extracting from issue body
+  const issue = await github.getIssue();
+  if (!issue)
+    throw new Error(
+      "No issue found in the context and no videoUrl provided. This action requires either an issue to be present or a videoUrl parameter.",
+    );
 
-const RX = /^https:\/\/github.com\/user-attachments\/assets\/.+$/gim;
-const assetLinks = Array.from(
-  new Set(Array.from(issue.body.matchAll(RX), (m) => m[0])),
-);
-if (assetLinks.length === 0)
-  cancel("No video assets found in the issue body, nothing to do.");
+  const RX = /^https:\/\/github.com\/user-attachments\/assets\/.+$/gim;
+  const assetLinks = Array.from(
+    new Set(Array.from(issue.body.matchAll(RX), (m) => m[0])),
+  );
+  if (assetLinks.length === 0)
+    cancel("No video assets found in the issue body, nothing to do.");
 
-dbg(`issue: %s`, issue.title);
+  dbg(`issue: %s`, issue.title);
 
-for (const assetLink of assetLinks) await processAssetLink(assetLink);
+  for (const assetLink of assetLinks) await processAssetLink(assetLink);
+}
 
 async function processAssetLink(assetLink: string) {
-  output.heading(3, assetLink);
+  output.heading(4, assetLink);
   dbg(assetLink);
   const downloadUrl = await github.resolveAssetUrl(assetLink);
   const res = await fetch(downloadUrl, { method: "GET" });
@@ -75,7 +87,7 @@ async function processVideo(filename: string) {
     (ctx) => {
       ctx.def("TRANSCRIPT", transcript?.srt, { ignoreEmpty: true }); // ignore silent videos
       ctx.defImages(frames, { detail: "low", sliceSample: 40 }); // low detail for better performance
-      ctx.$`${instructions}
+      ctx.$`${finalInstructions}
 ## Output format
 - Use GitHub Flavored Markdown (GFM) for markdown syntax formatting.
 - If you need to list tasks, use the format \`- [ ] task description\`.
@@ -99,3 +111,32 @@ async function processVideo(filename: string) {
     output.appendContent(text);
   }
 }
+
+async function processDirectVideoUrl(videoUrl: string) {
+  output.heading(4, videoUrl);
+  dbg(`Processing direct video URL: ${videoUrl}`);
+  
+  // Download video from direct URL
+  const res = await fetch(videoUrl, { method: "GET" });
+  const contentType = res.headers.get("content-type") || "";
+  dbg(`download url: %s`, videoUrl);
+  dbg(`headers: %O`, res.headers);
+  
+  if (!res.ok)
+    throw new Error(
+      `Failed to download video from ${videoUrl}: ${res.status} ${res.statusText}`,
+    );
+  
+  if (!/^video\//.test(contentType)) {
+    output.p(`URL does not point to a video file, skipping`);
+    return;
+  }
+
+  // save and cache
+  const buffer = await res.arrayBuffer();
+  dbg(`size`, `${(buffer.byteLength / 1e6) | 0}Mb`);
+  const filename = await workspace.writeCached(buffer, { scope: "run" });
+  dbg(`filename`, filename);
+
+  await processVideo(filename);
+}