adding huggingface inference support + litellm debugger

INSTALL.md

@@ -135,6 +135,14 @@ python pr_agent/cli.py --pr_url <pr_url> describe
 python pr_agent/cli.py --pr_url <pr_url> improve
 ```
 
+5. **Debugging LLM API Calls**  
+If you're testing your codium/pr-agent server, and need to see if calls were made successfully + the exact call logs, you can use the [LiteLLM Debugger tool](https://docs.litellm.ai/docs/debugging/hosted_debugging). 
+
+You can do this by setting `litellm_debugger=true` in configuration.toml. Your Logs will be viewable in real-time @ `admin.litellm.ai/<your_email>`. Set your email in the `.secrets.toml` under 'user_email'.
+
+<img src="./pics/debugger.png" width="900"/>
+
+
 ---
 
 #### Method 4: Run as a polling server

pr_agent/algo/ai_handler.py

@@ -26,6 +26,7 @@ class AiHandler:
         try:
             openai.api_key = get_settings().openai.key
             litellm.openai_key = get_settings().openai.key
+            litellm.debugger = get_settings().config.litellm_debugger
             self.azure = False
             if get_settings().get("OPENAI.ORG", None):
                 litellm.organization = get_settings().openai.org
@@ -43,6 +44,12 @@ class AiHandler:
                 litellm.cohere_key = get_settings().cohere.key
             if get_settings().get("REPLICATE.KEY", None):
                 litellm.replicate_key = get_settings().replicate.key
+            if get_settings().get("REPLICATE.KEY", None):
+                litellm.replicate_key = get_settings().replicate.key
+            if get_settings().get("HUGGINGFACE.KEY", None):
+                litellm.huggingface_key = get_settings().huggingface.key
+            if get_settings().get("HUGGINGFACE.KEY", None):
+                litellm.huggingface_key = get_settings().huggingface.key
         except AttributeError as e:
             raise ValueError("OpenAI key is required") from e
 
@@ -82,6 +89,8 @@ class AiHandler:
                     f"Generating completion with {model}"
                     f"{(' from deployment ' + deployment_id) if deployment_id else ''}"
                 )
+            if self.azure:
+                model = self.azure + "/" + model
             response = await acompletion(
                 model=model,
                 deployment_id=deployment_id,
@@ -90,7 +99,6 @@ class AiHandler:
                     {"role": "user", "content": user}
                 ],
                 temperature=temperature,
-                azure=self.azure,
                 force_timeout=get_settings().config.ai_timeout
             )
         except (APIError, Timeout, TryAgain) as e:

pr_agent/settings/configuration.toml

@@ -10,6 +10,7 @@ use_repo_settings_file=true
 ai_timeout=180
 max_description_tokens = 500
 max_commits_tokens = 500
+litellm_debugger=false
 
 [pr_reviewer] # /review #
 require_focused_review=true

requirements.txt

@@ -14,4 +14,4 @@ GitPython~=3.1.32
 litellm~=0.1.351
 PyYAML==6.0
 starlette-context==0.3.6
-litellm~=0.1.351
+litellm~=0.1.445