Merge pull request #786 from danieldekay/improved_azure_openai_config

Improved azure openai config
assafelovic · Aug 21, 2024 · e0e8524 · e0e8524
2 parents ffa9c94 + bb87592
commit e0e8524
Show file tree

Hide file tree

Showing 4 changed files with 64 additions and 31 deletions.
diff --git a/docs/docs/gpt-researcher/llms.md b/docs/docs/gpt-researcher/llms.md
@@ -46,9 +46,18 @@ OPENAI_EMBEDDING_MODEL="custom_model"
 
 ### Azure OpenAI
 
+See also the documentation in the Langchain [Azure OpenAI](https://api.python.langchain.com/en/latest/chat_models/langchain_openai.chat_models.azure.AzureChatOpenAI.html) page.
+
+On Azure OpenAI you will need to create deployments for each model you want to use. Please also specify the model names/deployment names in your `.env` file:
+
 ```bash
 EMBEDDING_PROVIDER="azure_openai"
 AZURE_OPENAI_API_KEY="Your key"
+AZURE_OPENAI_ENDPOINT="https://<your-endpoint>.openai.azure.com/"
+OPENAI_API_VERSION="2024-05-01-preview"
+FAST_LLM_MODEL="gpt-4o-mini"
+DEFAULT_LLM_MODEL="gpt-4o-mini"
+SMART_LLM_MODEL="gpt-4o"
 ```
 
 

diff --git a/gpt_researcher/config/config.py b/gpt_researcher/config/config.py
@@ -8,34 +8,43 @@ class Config:
 
     def __init__(self, config_file: str = None):
         """Initialize the config class."""
-        self.config_file = os.path.expanduser(config_file) if config_file else os.getenv('CONFIG_FILE')
-        self.retrievers = self.parse_retrievers(os.getenv('RETRIEVER', "tavily"))
-        self.embedding_provider = os.getenv('EMBEDDING_PROVIDER', 'openai')
-        self.similarity_threshold = int(os.getenv('SIMILARITY_THRESHOLD', 0.38))
-        self.llm_provider = os.getenv('LLM_PROVIDER', "openai")
-        self.ollama_base_url = os.getenv('OLLAMA_BASE_URL', None)
-        self.llm_model = "gpt-4o-mini"  
-        self.fast_llm_model = os.getenv('FAST_LLM_MODEL', "gpt-4o-mini")
-        self.smart_llm_model = os.getenv('SMART_LLM_MODEL', "gpt-4o-2024-08-06")
-        self.fast_token_limit = int(os.getenv('FAST_TOKEN_LIMIT', 2000))
-        self.smart_token_limit = int(os.getenv('SMART_TOKEN_LIMIT', 4000))
-        self.browse_chunk_max_length = int(os.getenv('BROWSE_CHUNK_MAX_LENGTH', 8192))
-        self.summary_token_limit = int(os.getenv('SUMMARY_TOKEN_LIMIT', 700))
-        self.temperature = float(os.getenv('TEMPERATURE', 0.55))
-        self.llm_temperature = float(os.getenv('LLM_TEMPERATURE', 0.55))  # Add this line
-        self.user_agent = os.getenv('USER_AGENT', "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
-                                                   "(KHTML, like Gecko) Chrome/119.0.0.0 Safari/537.36 Edg/119.0.0.0")
-        self.max_search_results_per_query = int(os.getenv('MAX_SEARCH_RESULTS_PER_QUERY', 5))
-        self.memory_backend = os.getenv('MEMORY_BACKEND', "local")
-        self.total_words = int(os.getenv('TOTAL_WORDS', 1000))
-        self.report_format = os.getenv('REPORT_FORMAT', "APA")
-        self.max_iterations = int(os.getenv('MAX_ITERATIONS', 3))
-        self.agent_role = os.getenv('AGENT_ROLE', None)
+        self.config_file = (
+            os.path.expanduser(config_file) if config_file else os.getenv("CONFIG_FILE")
+        )
+        self.retrievers = self.parse_retrievers(os.getenv("RETRIEVER", "tavily"))
+        self.embedding_provider = os.getenv("EMBEDDING_PROVIDER", "openai")
+        self.similarity_threshold = int(os.getenv("SIMILARITY_THRESHOLD", 0.38))
+        self.llm_provider = os.getenv("LLM_PROVIDER", "openai")
+        self.ollama_base_url = os.getenv("OLLAMA_BASE_URL", None)
+        self.llm_model = os.getenv("DEFAULT_LLM_MODEL", "gpt-4o-mini")
+        self.fast_llm_model = os.getenv("FAST_LLM_MODEL", "gpt-4o-mini")
+        self.smart_llm_model = os.getenv("SMART_LLM_MODEL", "gpt-4o-2024-08-06")
+        self.fast_token_limit = int(os.getenv("FAST_TOKEN_LIMIT", 2000))
+        self.smart_token_limit = int(os.getenv("SMART_TOKEN_LIMIT", 4000))
+        self.browse_chunk_max_length = int(os.getenv("BROWSE_CHUNK_MAX_LENGTH", 8192))
+        self.summary_token_limit = int(os.getenv("SUMMARY_TOKEN_LIMIT", 700))
+        self.temperature = float(os.getenv("TEMPERATURE", 0.55))
+        self.llm_temperature = float(
+            os.getenv("LLM_TEMPERATURE", 0.55)
+        )  # Add this line
+        self.user_agent = os.getenv(
+            "USER_AGENT",
+            "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
+            "(KHTML, like Gecko) Chrome/119.0.0.0 Safari/537.36 Edg/119.0.0.0",
+        )
+        self.max_search_results_per_query = int(
+            os.getenv("MAX_SEARCH_RESULTS_PER_QUERY", 5)
+        )
+        self.memory_backend = os.getenv("MEMORY_BACKEND", "local")
+        self.total_words = int(os.getenv("TOTAL_WORDS", 1000))
+        self.report_format = os.getenv("REPORT_FORMAT", "APA")
+        self.max_iterations = int(os.getenv("MAX_ITERATIONS", 3))
+        self.agent_role = os.getenv("AGENT_ROLE", None)
         self.scraper = os.getenv("SCRAPER", "bs")
         self.max_subtopics = os.getenv("MAX_SUBTOPICS", 3)
         self.report_source = os.getenv("REPORT_SOURCE", None)
         self.doc_path = os.getenv("DOC_PATH", "")
-        self.llm_kwargs = {} 
+        self.llm_kwargs = {}
 
         self.load_config_file()
         if not hasattr(self, "llm_kwargs"):
@@ -47,14 +56,26 @@ def __init__(self, config_file: str = None):
     def parse_retrievers(self, retriever_str: str):
         """Parse the retriever string into a list of retrievers and validate them."""
         VALID_RETRIEVERS = [
-            "arxiv", "bing", "custom", "duckduckgo", "exa", "google", "searx",
-            "semantic_scholar", "serpapi", "serper", "tavily", "pubmed_central"
+            "arxiv",
+            "bing",
+            "custom",
+            "duckduckgo",
+            "exa",
+            "google",
+            "searx",
+            "semantic_scholar",
+            "serpapi",
+            "serper",
+            "tavily",
+            "pubmed_central",
         ]
-        retrievers = [retriever.strip() for retriever in retriever_str.split(',')]
+        retrievers = [retriever.strip() for retriever in retriever_str.split(",")]
         invalid_retrievers = [r for r in retrievers if r not in VALID_RETRIEVERS]
         if invalid_retrievers:
-            raise ValueError(f"Invalid retriever(s) found: {', '.join(invalid_retrievers)}. "
-                             f"Valid options are: {', '.join(VALID_RETRIEVERS)}.")
+            raise ValueError(
+                f"Invalid retriever(s) found: {', '.join(invalid_retrievers)}. "
+                f"Valid options are: {', '.join(VALID_RETRIEVERS)}."
+            )
         return retrievers
 
     def validate_doc_path(self):

diff --git a/main.py b/main.py
@@ -1,8 +1,10 @@
-from backend.server import app
 from dotenv import load_dotenv
+
 load_dotenv()
 
+from backend.server import app
+
 if __name__ == "__main__":
     import uvicorn
 
-    uvicorn.run(app, host="0.0.0.0", port=8000)
+    uvicorn.run(app, host="0.0.0.0", port=8000)
diff --git a/requirements.txt b/requirements.txt
@@ -12,6 +12,7 @@ markdown
 langchain>=0.2,<0.3
 langchain_community>=0.2,<0.3
 langchain-openai>=0.1,<0.2
+langgraph
 tiktoken
 gpt-researcher
 arxiv