feat: add support for qwen models

.github/workflows/open_code_review.yml

@@ -4,7 +4,9 @@ on:
   workflow_call:
     secrets:
       DEEPSEEK_API_KEY:
-        required: true
+        required: false
+      QWEN_API_KEY:
+        required: false
 
 # A newer authorized /review command supersedes a review that is still running.
 # Other comments get unique groups, so they cannot cancel a PR review.
@@ -35,46 +37,95 @@ jobs:
     runs-on: ubuntu-latest
     timeout-minutes: 30
     steps:
+      - name: Acknowledge review command
+        uses: actions/github-script@v7
+        continue-on-error: true
+        with:
+          script: |
+            await github.rest.reactions.createForIssueComment({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              comment_id: context.payload.comment.id,
+              content: 'eyes',
+            });
+
       - name: Resolve model and pull request refs
         id: pr
         uses: actions/github-script@v7
         with:
           script: |
-            const command = context.payload.comment.body.trim();
-            const match = /^\/review(?: +([A-Za-z0-9-]+))?$/.exec(command);
+            const providers = {
+              deepseek: {
+                url: 'https://api.deepseek.com/chat/completions',
+                secret: 'DEEPSEEK_API_KEY',
+              },
+              qwen: {
+                url: 'https://dashscope-us.aliyuncs.com/compatible-mode/v1/chat/completions',
+                secret: 'QWEN_API_KEY',
+              },
+            };
+
             const models = {
-              pro: 'deepseek-v4-pro',
-              flash: 'deepseek-flash',
-              'deepseek-v4-pro': 'deepseek-v4-pro',
-              'deepseek-flash': 'deepseek-flash',
+              pro: { id: 'deepseek-v4-pro', provider: 'deepseek' },
+              flash: { id: 'deepseek-flash', provider: 'deepseek' },
+              max: { id: 'qwen3.8-max', provider: 'qwen' },
+              // DeepSeek
+              'deepseek-pro': { id: 'deepseek-v4-pro', provider: 'deepseek' },
+              'deepseek-flash': { id: 'deepseek-flash', provider: 'deepseek' },
+              // Qwen
+              'qwen-max': { id: 'qwen3.8-max', provider: 'qwen' },
+              'qwen-flash': { id: 'qwen3.8-flash', provider: 'qwen' },
+              'qwen-plus': { id: 'qwen3.7-plus', provider: 'qwen' },
             };
-            const model = match && (match[1] ? models[match[1]] : models.pro);
-            if (!model) {
-              core.setFailed('Use /review [pro|flash|deepseek-v4-pro|deepseek-flash].');
-              return;
-            }
-            core.setOutput('model', model);
 
-            const { data: pr } = await github.rest.pulls.get({
-              owner: context.repo.owner,
-              repo: context.repo.repo,
-              pull_number: context.issue.number,
-            });
-            core.setOutput('base_ref', pr.base.ref);
-            core.setOutput('head_sha', pr.head.sha);
+            try {
+              const command = context.payload.comment.body.trim();
+              const match = /^\/review(?: +([A-Za-z0-9._-]+))?$/i.exec(command);
+              const requestedModel = match && (match[1] || 'pro').toLowerCase();
+              const model = requestedModel && models[requestedModel];
+              if (!model) {
+                throw new Error(
+                  `Unsupported model. Use /review MODEL. Supported models: ${Object.keys(models).join(', ')}.`
+                );
+              }
 
-      - name: Acknowledge review command
+              const provider = providers[model.provider];
+              if (!provider) {
+                throw new Error(`Provider configuration is missing for ${model.provider}.`);
+              }
+
+              const { data: pr } = await github.rest.pulls.get({
+                owner: context.repo.owner,
+                repo: context.repo.repo,
+                pull_number: context.issue.number,
+              });
+
+              core.setOutput('model', model.id);
+              core.setOutput('llm_url', provider.url);
+              core.setOutput('llm_auth_secret', provider.secret);
+              core.setOutput('base_ref', pr.base.ref);
+              core.setOutput('head_sha', pr.head.sha);
+            } catch (error) {
+              const message = error instanceof Error ? error.message : String(error);
+              core.setOutput('error', message);
+              core.setFailed(message);
+            }
+
+      - name: Report model resolution error
+        if: ${{ failure() && steps.pr.outcome == 'failure' }}
         uses: actions/github-script@v7
-        continue-on-error: true
+        env:
+          ERROR_MESSAGE: ${{ steps.pr.outputs.error || 'Failed to resolve the review model or pull request.' }}
         with:
           script: |
-            await github.rest.reactions.createForIssueComment({
+            await github.rest.issues.createComment({
               owner: context.repo.owner,
               repo: context.repo.repo,
-              comment_id: context.payload.comment.id,
-              content: 'eyes',
+              issue_number: context.issue.number,
+              body: `OpenCodeReview could not start: ${process.env.ERROR_MESSAGE}`,
             });
 
+
       # The shared rules are checked out from this trusted repository, never
       # from the pull request branch being reviewed.
       - name: Check out shared Nim rules
@@ -111,8 +162,8 @@ jobs:
       - name: Review with OpenCodeReview
         uses: alibaba/open-code-review@main
         with:
-          llm_url: https://api.deepseek.com/chat/completions
-          llm_auth_token: ${{ secrets.DEEPSEEK_API_KEY }}
+          llm_url: ${{ steps.pr.outputs.llm_url }}
+          llm_auth_token: ${{ secrets[steps.pr.outputs.llm_auth_secret] }}
           llm_model: ${{ steps.pr.outputs.model }}
           llm_use_anthropic: 'false'
           base_ref: ${{ steps.pr.outputs.base_ref }}