Get llm adapter

.github/workflows/saist.yml

@@ -4,6 +4,20 @@ on:
   pull_request:
     types: [opened, synchronize, reopened]
 
+  workflow_dispatch:
+    inputs:
+      llm_provider:
+        description: 'LLM Provider'
+        required: true
+        default: 'deepseek'
+        type: choice
+        options:
+          - openai
+          - anthropic
+          - deepseek
+          - gemini
+          - ollama
+
 jobs:
   security-check:
     runs-on: ubuntu-latest
@@ -15,6 +29,8 @@ jobs:
     steps:
       - name: Check out code
         uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
 
       - name: Set up Python
         uses: actions/setup-python@v4
@@ -25,13 +41,93 @@ jobs:
         run: |
           pip install -r requirements.txt
 
+      - name: Determine LLM Provider and Scan Mode
+        id: config
+        run: |
+          if [ "${{ github.event_name }}" = "workflow_dispatch" ]; then
+            LLM_PROVIDER="${{ github.event.inputs.llm_provider }}"
+            SCAN_MODE="filesystem"
+            SCAN_TARGET="."
+            echo "🎯 Manual run: Using $LLM_PROVIDER with filesystem scan"
+          else
+            # PR run - use default provider with proper fallback
+            if [ -n "${{ vars.DEFAULT_LLM_PROVIDER }}" ]; then
+              LLM_PROVIDER="${{ vars.DEFAULT_LLM_PROVIDER }}"
+            else
+              LLM_PROVIDER="deepseek"
+            fi
+            SCAN_MODE="github"
+            SCAN_TARGET=""
+            echo "🔄 PR run: Using $LLM_PROVIDER with github scan"
+          fi
+
+          # Save for next steps
+          echo "llm_provider=$LLM_PROVIDER" >> $GITHUB_OUTPUT
+          echo "scan_mode=$SCAN_MODE" >> $GITHUB_OUTPUT
+          echo "scan_target=$SCAN_TARGET" >> $GITHUB_OUTPUT
+
+          # Set API key secret name
+          case $LLM_PROVIDER in
+            openai) echo "api_key_secret=OPENAI_API_KEY" >> $GITHUB_OUTPUT ;;
+            anthropic) echo "api_key_secret=ANTHROPIC_API_KEY" >> $GITHUB_OUTPUT ;;
+            deepseek) echo "api_key_secret=DEEPSEEK_API_KEY" >> $GITHUB_OUTPUT ;;
+            gemini) echo "api_key_secret=GEMINI_API_KEY" >> $GITHUB_OUTPUT ;;
+            ollama) echo "api_key_secret=NONE" >> $GITHUB_OUTPUT ;;
+          esac
+
       - name: Run Security Review
-        # We do NOT need a separate secret for GITHUB_TOKEN; it's provided by GitHub automatically.
-        # We DO need a secret for OPENAI_API_KEY, stored in your repo or org secrets.
         env:
-          GITHUB_TOKEN: ${{ github.token }}  # The pipeline token
+          GITHUB_TOKEN: ${{ github.token }}
           GITHUB_REPOSITORY: ${{ github.repository }}
           PR_NUMBER: ${{ github.event.pull_request.number }}
-          SAIST_LLM_API_KEY: ${{ secrets.DEEPSEEK_API_KEY }}
         run: |
-          python saist/main.py --llm deepseek github
+          echo "🔍 Running SAIST with ${{ steps.config.outputs.llm_provider }}"
+          echo "📁 Scan mode: ${{ steps.config.outputs.scan_mode }}"
+
+          # Get the API key for the selected provider
+          case "${{ steps.config.outputs.llm_provider }}" in
+            openai)
+              SAIST_LLM_API_KEY="${{ secrets.OPENAI_API_KEY }}"
+              ;;
+            anthropic)
+              SAIST_LLM_API_KEY="${{ secrets.ANTHROPIC_API_KEY }}"
+              ;;
+            deepseek)
+              SAIST_LLM_API_KEY="${{ secrets.DEEPSEEK_API_KEY }}"
+              ;;
+            gemini)
+              SAIST_LLM_API_KEY="${{ secrets.GEMINI_API_KEY }}"
+              ;;
+            ollama)
+              SAIST_LLM_API_KEY=""  # Ollama doesn't need an API key
+              ;;
+          esac
+
+          # Build the command with API key checking
+          if [ "${{ steps.config.outputs.scan_mode }}" = "github" ]; then
+            # PR mode - scan the github PR with token
+            if [ -n "$SAIST_LLM_API_KEY" ]; then
+              python saist/main.py --llm ${{ steps.config.outputs.llm_provider }} --llm-api-key "$SAIST_LLM_API_KEY" github ${{ github.repository }} ${{ github.event.pull_request.number }} --github-token "${{ github.token }}"
+            else
+              python saist/main.py --llm ${{ steps.config.outputs.llm_provider }} github ${{ github.repository }} ${{ github.event.pull_request.number }} --github-token "${{ github.token }}"
+            fi
+          else
+            # Manual mode - scan filesystem
+            if [ -n "$SAIST_LLM_API_KEY" ]; then
+              python saist/main.py --llm ${{ steps.config.outputs.llm_provider }} --llm-api-key "$SAIST_LLM_API_KEY" --csv filesystem ${{ steps.config.outputs.scan_target }}
+            else
+              python saist/main.py --llm ${{ steps.config.outputs.llm_provider }} --csv filesystem ${{ steps.config.outputs.scan_target }}
+            fi
+          fi
+
+          echo "✅ SAIST scan completed"
+
+      - name: Upload Results (Manual runs only)
+        if: github.event_name == 'workflow_dispatch' && always()
+        uses: actions/upload-artifact@v4
+        with:
+          name: saist-results-${{ github.run_number }}
+          path: |
+            findings.csv
+            *.pdf
+          retention-days: 30

saist/main.py

@@ -119,7 +119,7 @@ async def _get_llm_adapter(args) -> BaseLlmAdapter:
     if args.llm == 'anthropic':
         llm = AnthropicAdapter( api_key = args.llm_api_key, model=model)
         logger.debug(f"Using LLM: anthropic Model: {llm.model_name}")
-    if args.llm == 'bedrock':
+    elif args.llm == 'bedrock':
         llm = BedrockAdapter( api_key = args.llm_api_key, model=model)
         logger.debug(f"Using LLM: AWS bedrock Model: {llm.model_name}")
     elif args.llm == 'deepseek':