cat-llm 0.0.43__tar.gz → 0.0.51__tar.gz

{cat_llm-0.0.43 → cat_llm-0.0.51}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cat-llm
-Version: 0.0.43
+Version: 0.0.51
 Summary: A tool for categorizing text data and images using LLMs and vision models
 Project-URL: Documentation, https://github.com/chrissoria/cat-llm#readme
 Project-URL: Issues, https://github.com/chrissoria/cat-llm/issues

{cat_llm-0.0.43 → cat_llm-0.0.51}/src/catllm/CERAD_functions.py

@@ -374,7 +374,10 @@ def cerad_drawn_score(
     normalized_data = pd.concat(normalized_data_list, ignore_index=True)
 
     categorized_data = pd.DataFrame({
-        'image_input': image_files,
+        'image_input': (
+            image_files.reset_index(drop=True) if isinstance(image_files, (pd.DataFrame, pd.Series)) 
+            else pd.Series(image_files)
+        ),
         'link1': pd.Series(link1).reset_index(drop=True),
         'json': pd.Series(extracted_jsons).reset_index(drop=True)
     })

{cat_llm-0.0.43 → cat_llm-0.0.51}/src/catllm/__about__.py

@@ -1,7 +1,7 @@
 # SPDX-FileCopyrightText: 2025-present Christopher Soria <chrissoria@berkeley.edu>
 #
 # SPDX-License-Identifier: MIT
-__version__ = "0.0.43"
+__version__ = "0.0.51"
 __author__ = "Chris Soria"
 __email__ = "chrissoria@berkeley.edu"
 __title__ = "cat-llm"

{cat_llm-0.0.43 → cat_llm-0.0.51}/src/catllm/__init__.py

@@ -13,4 +13,5 @@ from .__about__ import (
 
 from .text_functions import *
 from .CERAD_functions import * 
-from .image_functions import * 
+from .image_functions import * 
+from .build_web_research import * 

cat_llm-0.0.51/src/catllm/build_web_research.py

@@ -0,0 +1,169 @@
+#build dataset classification
+def build_web_research_dataset(
+    search_question, 
+    search_input,
+    api_key,
+    answer_format = "concise",
+    additional_instructions = "",
+    categories = ['Answer','URL'],
+    user_model="claude-3-7-sonnet-20250219",
+    creativity=0,
+    safety=False,
+    filename="categorized_data.csv",
+    save_directory=None,
+    model_source="Anthropic",
+    time_delay=5
+):
+    import os
+    import json
+    import pandas as pd
+    import regex
+    from tqdm import tqdm
+    import time
+    
+    categories_str = "\n".join(f"{i + 1}. {cat}" for i, cat in enumerate(categories))
+    print(categories_str)
+    cat_num = len(categories)
+    category_dict = {str(i+1): "0" for i in range(cat_num)}
+    example_JSON = json.dumps(category_dict, indent=4)
+
+    # ensure number of categories is what user wants
+    #print("\nThe information to be extracted:")
+    #for i, cat in enumerate(categories, 1):
+        #print(f"{i}. {cat}")
+    
+    link1 = []
+    extracted_jsons = []
+
+    for idx, item in enumerate(tqdm(search_input, desc="Building dataset")):
+        if idx == 0:  # delay the first item just to be safe
+            time.sleep(time_delay)
+        reply = None  
+
+        if pd.isna(item): 
+            link1.append("Skipped NaN input")
+            default_json = example_JSON 
+            extracted_jsons.append(default_json)
+            #print(f"Skipped NaN input.")
+        else:
+            prompt = f"""<role>You are a research assistant specializing in finding current, factual information.</role>
+
+            <task>Find information about {item}'s {search_question}</task>
+
+            <rules>
+            - Search for the most current and authoritative information available
+            - Provide your answer as {answer_format}
+            - Prioritize official sources when possible
+            - If information is not found, state "Information not found"
+            - Include exactly one source URL where you found the information
+            - Do not include any explanatory text or commentary beyond the JSON
+                {additional_instructions}
+            </rules>
+
+            <format>
+            Return your response as valid JSON with this exact structure:
+            {{
+            "answer": "Your factual answer or 'Information not found'",
+            "url": "Source URL or 'No source available'"
+        }}
+        </format>"""
+            #print(prompt)
+            if model_source == "Anthropic":
+                import anthropic
+                client = anthropic.Anthropic(api_key=api_key)
+                try:
+                    message = client.messages.create(
+                    model=user_model,
+                    max_tokens=1024,
+                    temperature=creativity,
+                    messages=[{"role": "user", "content": prompt}],
+                    tools=[{
+                    "type": "web_search_20250305", 
+                    "name": "web_search"
+                    }]
+                )
+                    reply = " ".join(
+                        block.text
+                        for block in message.content
+                        if getattr(block, "type", "") == "text"
+                    ).strip()
+                    link1.append(reply)
+                    time.sleep(time_delay)
+                    print(reply)
+                    
+                except Exception as e:
+                    print(f"An error occurred: {e}")
+                    link1.append(f"Error processing input: {e}")
+                    time.sleep(time_delay)
+            else:
+                raise ValueError("Unknown source! Currently this function only supports 'Anthropic' as model_source.")
+            # in situation that no JSON is found
+            if reply is not None:
+                extracted_json = regex.findall(r'\{(?:[^{}]|(?R))*\}', reply, regex.DOTALL)
+                if extracted_json:
+                    raw_json = extracted_json[0].strip()  # Only strip leading/trailing whitespace
+                    try:
+                        # Parse to validate JSON structure
+                        parsed_obj = json.loads(raw_json)
+                        # Re-serialize for consistent formatting (optional)
+                        cleaned_json = json.dumps(parsed_obj)
+                        extracted_jsons.append(cleaned_json)
+                    except json.JSONDecodeError as e:
+                        print(f"JSON parsing error: {e}")
+                        # Fallback to raw extraction if parsing fails
+                        extracted_jsons.append(raw_json)
+                else:
+                    # Use consistent schema for errors
+                    error_message = json.dumps({"answer": "e", "url": "e"})
+                    extracted_jsons.append(error_message)
+                    print(error_message)
+            else:
+                # Handle None reply case
+                error_message = json.dumps({"answer": "e", "url": "e"})
+                extracted_jsons.append(error_message)
+                #print(error_message)
+
+        # --- Safety Save ---
+        if safety:
+            # Save progress so far
+            temp_df = pd.DataFrame({
+                'survey_response': search_input[:idx+1],
+                'link1': link1,
+                'json': extracted_jsons
+            })
+            # Normalize processed jsons so far
+            normalized_data_list = []
+            for json_str in extracted_jsons:
+                try:
+                    parsed_obj = json.loads(json_str)
+                    normalized_data_list.append(pd.json_normalize(parsed_obj))
+                except json.JSONDecodeError:
+                    normalized_data_list.append(pd.DataFrame({"1": ["e"]}))
+            normalized_data = pd.concat(normalized_data_list, ignore_index=True)
+            temp_df = pd.concat([temp_df, normalized_data], axis=1)
+            # Save to CSV
+            if save_directory is None:
+                save_directory = os.getcwd()
+            temp_df.to_csv(os.path.join(save_directory, filename), index=False)
+
+    # --- Final DataFrame ---
+    normalized_data_list = []
+    for json_str in extracted_jsons:
+        try:
+            parsed_obj = json.loads(json_str)
+            normalized_data_list.append(pd.json_normalize(parsed_obj))
+        except json.JSONDecodeError:
+            normalized_data_list.append(pd.DataFrame({"1": ["e"]}))
+    normalized_data = pd.concat(normalized_data_list, ignore_index=True)
+
+    categorized_data = pd.DataFrame({
+        'survey_response': (
+            search_input.reset_index(drop=True) if isinstance(search_input, (pd.DataFrame, pd.Series)) 
+            else pd.Series(search_input)
+        ),
+        'link1': pd.Series(link1).reset_index(drop=True),
+        'json': pd.Series(extracted_jsons).reset_index(drop=True)
+    })
+    categorized_data = pd.concat([categorized_data, normalized_data], axis=1)
+    
+    return categorized_data

{cat_llm-0.0.43 → cat_llm-0.0.51}/src/catllm/image_functions.py

@@ -252,9 +252,11 @@ def image_multi_class(
         except json.JSONDecodeError:
             normalized_data_list.append(pd.DataFrame({"1": ["e"]}))
     normalized_data = pd.concat(normalized_data_list, ignore_index=True)
-
     categorized_data = pd.DataFrame({
-        'image_input': image_files,
+        'image_input': (
+            image_files.reset_index(drop=True) if isinstance(image_files, (pd.DataFrame, pd.Series)) 
+            else pd.Series(image_files)
+        ),
         'link1': pd.Series(link1).reset_index(drop=True),
         'json': pd.Series(extracted_jsons).reset_index(drop=True)
     })
@@ -549,7 +551,10 @@ def image_score_drawing(
     normalized_data = pd.concat(normalized_data_list, ignore_index=True)
 
     categorized_data = pd.DataFrame({
-        'image_input': image_files,
+        'image_input': (
+            image_files.reset_index(drop=True) if isinstance(image_files, (pd.DataFrame, pd.Series)) 
+            else pd.Series(image_files)
+        ),
         'link1': pd.Series(link1).reset_index(drop=True),
         'json': pd.Series(extracted_jsons).reset_index(drop=True)
     })
@@ -835,7 +840,10 @@ def image_features(
     normalized_data = pd.concat(normalized_data_list, ignore_index=True)
 
     categorized_data = pd.DataFrame({
-        'image_input': image_files,
+        'image_input': (
+            image_files.reset_index(drop=True) if isinstance(image_files, (pd.DataFrame, pd.Series)) 
+            else pd.Series(image_files)
+        ),
         'link1': pd.Series(link1).reset_index(drop=True),
         'json': pd.Series(extracted_jsons).reset_index(drop=True)
     })

{cat_llm-0.0.43 → cat_llm-0.0.51}/src/catllm/text_functions.py

@@ -373,20 +373,14 @@ Provide your work in JSON format where the number belonging to each category is
         except json.JSONDecodeError:
             normalized_data_list.append(pd.DataFrame({"1": ["e"]}))
     normalized_data = pd.concat(normalized_data_list, ignore_index=True)
-
     categorized_data = pd.DataFrame({
-        'survey_response': survey_input.reset_index(drop=True),
+        'image_input': (
+            survey_input.reset_index(drop=True) if isinstance(survey_input, (pd.DataFrame, pd.Series)) 
+            else pd.Series(survey_input)
+        ),
         'link1': pd.Series(link1).reset_index(drop=True),
         'json': pd.Series(extracted_jsons).reset_index(drop=True)
     })
     categorized_data = pd.concat([categorized_data, normalized_data], axis=1)
     
-    if columns != "numbered": #if user wants text columns
-        categorized_data.columns = list(categorized_data.columns[:3]) + categories[:len(categorized_data.columns) - 3]
-
-    if to_csv:
-        if save_directory is None:
-            save_directory = os.getcwd()
-        categorized_data.to_csv(os.path.join(save_directory, filename), index=False)
-    
     return categorized_data