From 8a08b0c01b5be1b70e58a1d6817d4f85c54465ef Mon Sep 17 00:00:00 2001
From: Preston Jensen <prestonjensen@Prestons-MacBook-Air.local>
Date: Mon, 3 Apr 2023 19:20:42 -0600
Subject: [PATCH] ask questions on websites

---
 scripts/browse.py       | 53 +++++++++++++----------------------------
 scripts/commands.py     | 12 +++++-----
 scripts/data/prompt.txt |  2 +-
 3 files changed, 24 insertions(+), 43 deletions(-)

diff --git a/scripts/browse.py b/scripts/browse.py
index f096c5f3e..510f9c29d 100644
--- a/scripts/browse.py
+++ b/scripts/browse.py
@@ -74,30 +74,25 @@ def split_text(text, max_length=8192):
         yield "\n".join(current_chunk)
 
 
-def summarize_text(text, is_website=True):
-    if text == "":
+def create_message(chunk, question):
+    return {
+        "role": "user",
+        "content": f"\"\"\"{chunk}\"\"\" Using the above text, please answer the following question: \"{question}\" -- if the question cannot be answered using the text, please summarize the text."
+    }
+
+def summarize_text(text, question):
+    if not text:
         return "Error: No text to summarize"
 
-    print("Text length: " + str(len(text)) + " characters")
+    text_length = len(text)
+    print(f"Text length: {text_length} characters")
+
     summaries = []
     chunks = list(split_text(text))
 
     for i, chunk in enumerate(chunks):
-        print("Summarizing chunk " + str(i + 1) + " / " + str(len(chunks)))
-        if is_website:
-            messages = [
-                {
-                    "role": "user",
-                    "content": "Please summarize the following website text, do not describe the general website, but instead concisely extract the specific information this subpage contains.: " +
-                    chunk},
-            ]
-        else:
-            messages = [
-                {
-                    "role": "user",
-                    "content": "Please summarize the following text, focusing on extracting concise and specific information: " +
-                    chunk},
-            ]
+        print(f"Summarizing chunk {i + 1} / {len(chunks)}")
+        messages = [create_message(chunk, question)]
 
         summary = create_chat_completion(
             model=cfg.fast_llm_model,
@@ -105,25 +100,11 @@ def summarize_text(text, is_website=True):
             max_tokens=300,
         )
         summaries.append(summary)
-    print("Summarized " + str(len(chunks)) + " chunks.")
+
+    print(f"Summarized {len(chunks)} chunks.")
 
     combined_summary = "\n".join(summaries)
-
-    # Summarize the combined summary
-    if is_website:
-        messages = [
-            {
-                "role": "user",
-                "content": "Please summarize the following website text, do not describe the general website, but instead concisely extract the specific information this subpage contains.: " +
-                combined_summary},
-        ]
-    else:
-        messages = [
-            {
-                "role": "user",
-                "content": "Please summarize the following text, focusing on extracting concise and specific infomation: " +
-                combined_summary},
-        ]
+    messages = [create_message(combined_summary, question)]
 
     final_summary = create_chat_completion(
         model=cfg.fast_llm_model,
@@ -131,4 +112,4 @@ def summarize_text(text, is_website=True):
         max_tokens=300,
     )
 
-    return final_summary
+    return final_summary
\ No newline at end of file
diff --git a/scripts/commands.py b/scripts/commands.py
index 8ad953364..ed789c8e9 100644
--- a/scripts/commands.py
+++ b/scripts/commands.py
@@ -72,7 +72,7 @@ def execute_command(command_name, arguments):
         elif command_name == "delete_agent":
             return delete_agent(arguments["key"])
         elif command_name == "get_text_summary":
-            return get_text_summary(arguments["url"])
+            return get_text_summary(arguments["url"], arguments["question"])
         elif command_name == "get_hyperlinks":
             return get_hyperlinks(arguments["url"])
         elif command_name == "read_file":
@@ -84,7 +84,7 @@ def execute_command(command_name, arguments):
         elif command_name == "delete_file":
             return delete_file(arguments["file"])
         elif command_name == "browse_website":
-            return browse_website(arguments["url"])
+            return browse_website(arguments["url"], arguments["question"])
         # TODO: Change these to take in a file rather than pasted code, if
         # non-file is given, return instructions "Input should be a python
         # filepath, write your code to file and try again"
@@ -152,8 +152,8 @@ def google_official_search(query, num_results=8):
     # Return the list of search result URLs
     return search_results_links
 
-def browse_website(url):
-    summary = get_text_summary(url)
+def browse_website(url, question):
+    summary = get_text_summary(url, question)
     links = get_hyperlinks(url)
 
     # Limit links to 5
@@ -165,9 +165,9 @@ def browse_website(url):
     return result
 
 
-def get_text_summary(url):
+def get_text_summary(url, question):
     text = browse.scrape_text(url)
-    summary = browse.summarize_text(text)
+    summary = browse.summarize_text(text, question)
     return """ "Result" : """ + summary
 
 
diff --git a/scripts/data/prompt.txt b/scripts/data/prompt.txt
index d17fa27a1..1596c8105 100644
--- a/scripts/data/prompt.txt
+++ b/scripts/data/prompt.txt
@@ -9,7 +9,7 @@ COMMANDS:
 2. Memory Add: "memory_add", args: "string": "<string>"
 3. Memory Delete: "memory_del", args: "key": "<key>"
 4. Memory Overwrite: "memory_ovr", args: "key": "<key>", "string": "<string>"
-5. Browse Website: "browse_website", args: "url": "<url>"
+5. Browse Website: "browse_website", args: "url": "<url>", "question": "<what_you_want_to_find_on_website>"
 6. Start GPT Agent: "start_agent",  args: "name": <name>, "task": "<short_task_desc>", "prompt": "<prompt>"
 7. Message GPT Agent: "message_agent", args: "key": "<key>", "message": "<message>"
 8. List GPT Agents: "list_agents", args: ""