Upload folder using huggingface_hub

README.md

@@ -88,8 +88,7 @@ Given a query, the program will
 - use the top chunks as the context to ask an LLM to generate the answer
 - output the answer with the references
 
-Of course this flow is a very simplified version of the real AI search engines, but it is a good
-starting point to understand the basic concepts.
+Of course this flow is a very simplified version of the real AI search engines, but it is a good starting point to understand the basic concepts.
 
 One benefit is that we can manipulate the search function and output format.
 
@@ -112,9 +111,9 @@ the pipeline.
 
 # modify .env file to set the API keys or export them as environment variables as below
 
-# you can use the Google search API, if not set we provide a default search engine proxy for testing
-# % export SEARCH_API_KEY="your-google-search-api-key"
-# % export SEARCH_PROJECT_KEY="your-google-cx-key"
+# you need to set the Google search API
+% export SEARCH_API_KEY="your-google-search-api-key"
+% export SEARCH_PROJECT_KEY="your-google-cx-key"
 
 # right now we use OpenAI API, default using OpenAI
 # % export LLM_BASE_URL=https://api.openai.com/v1

ask.py

@@ -135,29 +135,13 @@ class Ask:
 
         self.search_api_url = os.environ.get("SEARCH_API_URL")
         if self.search_api_url is None:
-            self.search_api_key = os.environ.get("SEARCH_API_KEY")
-            if self.search_api_key:
-                self.search_api_url = "https://www.googleapis.com/customsearch/v1"
-                self.search_project_id = os.environ.get("SEARCH_PROJECT_KEY")
-                if self.search_project_id is None:
-                    err_msg += "SEARCH_PROJECT_KEY env variable not set while SEARCH_API_KEY is set.\n"
-            else:
-                self.logger.info("No SEARCH_API_URL or SEARCH_API_KEYenv variable set.")
-                self.logger.info(
-                    "Using the default proxy at https://svc.leettools.com:8098"
-                )
-                self.search_api_url = "https://svc.leettools.com:8098/customsearch/v1"
-                self.search_api_key = "dummy-search-api-key"
-                self.search_project_id = "dummy-search-project-id"
-        else:
-            self.search_api_key = os.environ.get("SEARCH_API_KEY")
-            if self.search_api_key is None:
-                err_msg += (
-                    f"SEARCH_API_KEY env variable not set for {self.search_api_url}.\n"
-                )
-            self.search_project_id = os.environ.get("SEARCH_PROJECT_KEY")
-            if self.search_project_id is None:
-                err_msg += f"SEARCH_PROJECT_KEY env variable not set for {self.search_api_url}.\n"
+            self.search_api_url = "https://www.googleapis.com/customsearch/v1"
+        self.search_api_key = os.environ.get("SEARCH_API_KEY")
+        self.search_project_id = os.environ.get("SEARCH_PROJECT_KEY")
+        if self.search_api_key is None:
+            err_msg += "SEARCH_API_KEY env variable is not set.\n"
+        if self.search_project_id is None:
+            err_msg += "SEARCH_PROJECT_KEY env variable is not set.\n"
 
         self.llm_base_url = os.environ.get("LLM_BASE_URL")
         if self.llm_base_url is None:

requirements.txt

@@ -4,8 +4,8 @@ numpy==1.26.4
 jinja2==3.1.3
 bs4==0.0.2
 python-dotenv==1.0.1
-openai==1.57.2
-duckdb==1.1.2
-gradio==5.3.0
-chonkie==0.1.2
-docling==2.5.2
+openai==1.87.0
+duckdb==1.3.0
+gradio==5.34.0
+chonkie==v1.0.10
+docling==2.36.0