Upload folder using huggingface_hub

.env.tpl

@@ -3,4 +3,8 @@ SEARCH_API_KEY=your-google-search-api-key
 SEARCH_PROJECT_KEY=your-google-cx-key
 
 # right now we use OpenAI API
-LLM_API_KEY=your-openai-api-key
+LLM_API_KEY=your-openai-api-key
+
+# Run and share Gradio UI
+RUN_GRADIO_UI=Faslse
+SHARE_GRADIO_UI=False

README.md

@@ -11,6 +11,14 @@ sdk_version: 5.3.0
 A single Python program to implement the search-extract-summarize flow, similar to AI search
 engines such as Perplexity.
 
+> [UPDATE]
+>
+> - 2024-10-22: add GradIO integation
+> - 2024-10-21: use DuckDB for the vector search and use API for embedding
+> - 2024-10-20: allow to specify a list of input urls
+> - 2024-10-18: output-language and output-length parameters for LLM
+> - 2024-10-18: date-restrict and target-site parameters for seach
+
 > [!NOTE]
 > Our main goal is to illustrate the basic concepts of AI search engines with the raw constructs.
 > Performance or scalability is not in the scope of this program.
@@ -64,17 +72,17 @@ Usage: ask.py [OPTIONS]
   Search web for the query and summarize the results
 
 Options:
-  -q, --query TEXT                Query to search  [required]
-  --url-list TEXT                 Instead of doing web search, scrape the
-                                  target URL list and answer the query based
-                                  on the content  [default:
-                                  instructions/links.txt]
+  --web-ui                        Launch the web interface
+  -q, --query TEXT                Query to search
   -d, --date-restrict INTEGER     Restrict search results to a specific date
                                   range, default is no restriction
   -s, --target-site TEXT          Restrict search results to a specific site,
                                   default is no restriction
   --output-language TEXT          Output language for the answer
   --output-length INTEGER         Output length for the answer
+  --url-list-file TEXT            Instead of doing web search, scrape the
+                                  target URL list and answer the query based
+                                  on the content
   -m, --model-name TEXT           Model name to use for inference
   -l, --log-level [DEBUG|INFO|WARNING|ERROR]
                                   Set the logging level  [default: INFO]
@@ -87,7 +95,12 @@ Options:
 - [OpenAI API](https://beta.openai.com/docs/api-reference/completions/create)
 - [Jinja2](https://jinja.palletsprojects.com/en/3.0.x/)
 - [bs4](https://www.crummy.com/software/BeautifulSoup/bs4/doc/)
-- [duckdb](https://github.com/duckdb/duckdb)
+- [DuckDB](https://github.com/duckdb/duckdb)
+- [GradIO](https://grad.io)
+
+## Screenshot for the GradIO integration
+
+![image](https://github.com/user-attachments/assets/0483e6a2-75d7-4fbd-813f-bfa13839c836)
 
 ## Sample output
 

ask.py

@@ -410,8 +410,6 @@ def _run_query(
 ) -> str:
     logger = get_logger(log_level)
 
-    load_dotenv(dotenv_path=default_env_file, override=False)
-
     ask = Ask(logger=logger)
 
     if url_list_str is None or url_list_str.strip() == "":
@@ -474,6 +472,7 @@ def launch_gradio(
     url_list_str: str,
     model_name: str,
     log_level: str,
+    share_ui: bool,
 ) -> None:
     iface = gr.Interface(
         fn=_run_query,
@@ -513,7 +512,7 @@ def launch_gradio(
         description="Search the web with the query and summarize the results. Source code: https://github.com/pengfeng/ask.py",
     )
 
-    iface.launch()
+    iface.launch(share=share_ui)
 
 
 @click.command(help="Search web for the query and summarize the results")
@@ -586,7 +585,13 @@ def search_extract_summarize(
     model_name: str,
     log_level: str,
 ):
-    if web_ui:
+    load_dotenv(dotenv_path=default_env_file, override=False)
+
+    if web_ui or os.environ.get("RUN_GRADIO_UI", "false").lower() != "false":
+        if os.environ.get("SHARE_GRADIO_UI", "false").lower() == "true":
+            share_ui = True
+        else:
+            share_ui = False
         launch_gradio(
             query=query,
             date_restrict=date_restrict,
@@ -596,6 +601,7 @@ def search_extract_summarize(
             url_list_str=_read_url_list(url_list_file),
             model_name=model_name,
             log_level=log_level,
+            share_ui=share_ui,
         )
     else:
         if query is None: