feat: Added an openai api request that extracts the most relevant inf…

…ormation from each website in order to estimate the prediction market question later. This strategy replaced the strategy where only the first 300 characters from each website's html text were extracted and used for prediction.

tools/prediction_sum_url_content.py

@@ -32,7 +32,7 @@
 NUM_URLS_EXTRACT = 5
 DEFAULT_OPENAI_SETTINGS = {
     "max_tokens": 500,
-    "temperature": 0,
+    "temperature": 0.2,
 }
 ALLOWED_TOOLS = [
     "prediction-offline-sum-url-content",
@@ -108,9 +108,38 @@
    - "queries": An array of strings of size between 1 and 5. Each string must be a search engine query that can help obtain relevant information to estimate
      the probability that the event in "USER_PROMPT" occurs. You must provide original information in each query, and they should not overlap
      or lead to obtain the same set of results.
-* Output only the JSON object. Do not include any other contents in your response.
+* Output only the JSON object. Do not include any other contents in your response. 
 """
 
+SUMMARY_SYSTEM_PROMPT = """
+You are an LLM inside a multi-agent system that takes in a prompt of a user requesting a probability estimation
+for a given event. You are provided with input under the label "USER_PROMPT" and "WEBSITE_TEXT". You must follow the instructions
+under the label "INSTRUCTIONS". You must provide your response in the format specified under "OUTPUT_FORMAT".
+
+INSTRUCTIONS
+* Read the input under the label "USER_PROMPT" and "WEBSITE_TEXT", each delimited by three backticks.
+* You must extract the content inside "WEBSITE_TEXT" that can be used to estimate the outcome of the event described inside "USER_PROMPT".
+* You must provide your response in the format specified under "OUTPUT_FORMAT".
+* Do not include any other contents in your response except for those extracted from "WEBSITE_TEXT".
+
+USER_PROMPT:
+```
+{user_prompt}
+```
+
+WEBSITE_TEXT:
+```
+{website_text}
+```
+
+OUTPUT_FORMAT
+* Your output response must be only one string containing the most relevant statements, separated by a ".".
+* Provide only the extracted, relevant information for estimating the outcome of the event. 
+* Do not include any headers or introductory phrases.
+* Your response must not exceed 100 words.
+* If the content in "WEBSITE_TEXT" is not relevant for estimating the outcome of the event described in "USER_PROMPT", your response must be an empty string.
+
+"""
 
 def search_google(query: str, api_key: str, engine: str, num: int = 3) -> List[str]:
     service = build("customsearch", "v1", developerKey=api_key)
@@ -141,9 +170,39 @@ def get_urls_from_queries(queries: List[str], api_key: str, engine: str) -> List
     return unique_results
 
 
+def get_website_summary(
+    text: str,
+    prompt: str,
+    engine: str,
+    temperature: float,
+    max_tokens: int,
+) -> str:
+    """Get text summary from a website"""
+    user_prompt_summary = SUMMARY_SYSTEM_PROMPT.format(user_prompt=prompt, website_text=text)
+
+    messages = [
+        {"role": "system", "content": "You are a helpful assistant."},
+        {"role": "user", "content": user_prompt_summary},
+    ]
+    response = openai.ChatCompletion.create(
+        model=engine,
+        messages=messages,
+        temperature=temperature,
+        max_tokens=max_tokens,
+        n=1,
+        timeout=150,
+        request_timeout=150,
+        stop=None,
+    )
+    return response.choices[0].message.content
+
+
 def extract_text(
     html: str,
-    num_words: int = 300,  # TODO: summerise using GPT instead of limit
+    prompt: str,
+    engine: str,
+    temperature: float,
+    max_tokens: int,
 ) -> str:
     """Extract text from a single HTML document"""
     soup = BeautifulSoup(html, "html.parser")
@@ -153,7 +212,14 @@ def extract_text(
     lines = (line.strip() for line in text.splitlines())
     chunks = (phrase.strip() for line in lines for phrase in line.split("  "))
     text = "\n".join(chunk for chunk in chunks if chunk)
-    return text[:num_words]
+    text_summary = get_website_summary(
+        text=text,
+        prompt=prompt,
+        engine=engine,
+        temperature=temperature,
+        max_tokens=max_tokens,
+    )
+    return text_summary
 
 
 def process_in_batches(
@@ -166,7 +232,13 @@ def process_in_batches(
             futures = [(executor.submit(requests.get, url, timeout=timeout), url) for url in batch]
             yield futures
 
-def extract_texts(urls: List[str], num_words: int = 300) -> List[str]:
+def extract_texts(
+    urls: List[str],
+    prompt: str,
+    engine: str,
+    temperature: float,
+    max_tokens: int,
+) -> List[str]:
     """Extract texts from URLs"""
     max_allowed = 5
     extracted_texts = []
@@ -178,7 +250,14 @@ def extract_texts(urls: List[str], num_words: int = 300) -> List[str]:
                 result = future.result()
                 if result.status_code != 200:
                     continue
-                extracted_texts.append(extract_text(html=result.text, num_words=num_words))
+                extracted_text = extract_text(
+                    html=result.text,
+                    prompt=prompt,
+                    engine=engine,
+                    temperature=temperature,
+                    max_tokens=max_tokens,
+                )
+                extracted_texts.append(extracted_text)
                 count += 1
                 if count >= max_allowed:
                     stop = True
@@ -227,35 +306,17 @@ def fetch_additional_information(
         engine=google_engine,
     )
     print(f"urls: {urls}\n")
-    texts = extract_texts(urls)
+    texts = extract_texts(
+        urls=urls,
+        prompt=prompt,
+        engine=engine,
+        temperature=temperature,
+        max_tokens=max_tokens,
+    )
     additional_informations = "\n".join(["- " + text for text in texts])
     print(f"additional_informations: {additional_informations}\n")
     return additional_informations
 
-# # To be adjusted
-# def get_website_summary(engine, temperature, max_tokens, prompt) -> Tuple[str, str]:
-#     """Get SME title and introduction"""
-#     market_question = SME_GENERATION_MARKET_PROMPT.format(question=prompt)
-#     system_prompt = SME_GENERATION_SYSTEM_PROMPT
-
-#     messages = [
-#         {"role": "system", "content": system_prompt},
-#         {"role": "user", "content": market_question},
-#     ]
-#     response = openai.ChatCompletion.create(
-#         model=engine,
-#         messages=messages,
-#         temperature=temperature,
-#         max_tokens=max_tokens,
-#         n=1,
-#         timeout=150,
-#         request_timeout=150,
-#         stop=None,
-#     )
-#     generated_sme_roles = response.choices[0].message.content
-#     sme = json.loads(generated_sme_roles)[0]
-#     return sme["sme"], sme["sme_introduction"]
-
 
 def run(**kwargs) -> Tuple[str, Optional[Dict[str, Any]]]:
     """Run the task"""