ragaai-catalyst 2.0.3__py3-none-any.whl → 2.0.5__py3-none-any.whl

ragaai_catalyst/__init__.py

@@ -5,5 +5,7 @@ from .utils import response_checker
 from .dataset import Dataset
 from .prompt_manager import PromptManager
 from .evaluation import Evaluation
+from .synthetic_data_generation import SyntheticDataGeneration
 
-__all__ = ["Experiment", "RagaAICatalyst", "Tracer", "PromptManager", "Evaluation"]
+
+__all__ = ["Experiment", "RagaAICatalyst", "Tracer", "PromptManager", "Evaluation","SyntheticDataGeneration"]

ragaai_catalyst/evaluation.py

@@ -80,7 +80,8 @@ class Evaluation:
         try:
             response = requests.get(
                 f'{self.base_url}/v1/llm/llm-metrics', 
-                headers=headers)
+                headers=headers,
+                timeout=self.timeout)
             response.raise_for_status()
             metric_names = [metric["name"] for metric in response.json()["data"]["metrics"]]
             return metric_names
@@ -111,7 +112,8 @@ class Evaluation:
             response = requests.post(
                 f'{self.base_url}/v1/llm/docs', 
                 headers=headers,
-                json=data)
+                json=data,
+                timeout=self.timeout)
             response.raise_for_status()
             if response.status_code == 200:
                 return response.json()["data"]["columns"]
@@ -128,27 +130,52 @@ class Evaluation:
         return {}
 
     def _get_variablename_from_dataset_schema(self, schemaName, metric_name):
+        # pdb.set_trace()
+        # print(schemaName)
         dataset_schema = self._get_dataset_schema()
         variableName = None
         for column in dataset_schema:
-            columnName = column["columnName"].split('_')[0]
+            columnName = column["columnType"]
             displayName = column["displayName"]
-            if columnName==schemaName.lower():
+            # print(columnName, displayName)
+            if "".join(columnName.split("_")).lower() == schemaName.lower():
                 variableName = displayName
+                break
+        return variableName
+        # print(variableName)
+        # if variableName:
+        #     return variableName
+        # else:
+        #     raise ValueError(f"'{schemaName}' column is required for {metric_name} metric evaluation, but not found in dataset")
+
+
+    def _get_variablename_from_user_schema_mapping(self, schemaName, metric_name, schema_mapping):
+        # pdb.set_trace()
+        user_dataset_schema = self._get_dataset_schema()
+        user_dataset_columns = [item["displayName"] for item in user_dataset_schema]
+        variableName = None
+        for key, val in schema_mapping.items():
+            if "".join(val.split("_")).lower()==schemaName:
+                if key in user_dataset_columns:
+                    variableName=key
+                else:
+                    raise ValueError(f"Column '{key}' is not present in {self.dataset_name}")
         if variableName:
             return variableName
         else:
-            raise ValueError(f"'{schemaName.lower()}' column is required for {metric_name} metric evaluation, but not found in dataset")
+            raise ValueError(f"Map '{schemaName}' column in schema_mapping for {metric_name} metric evaluation")
 
 
-    def _get_mapping(self, metric_name, metrics_schema):
+    def _get_mapping(self, metric_name, metrics_schema, schema_mapping):
+        
         mapping = []
         for schema in metrics_schema:
             if schema["name"]==metric_name:
                 requiredFields = schema["config"]["requiredFields"]
                 for field in requiredFields:
                     schemaName = field["name"]
-                    variableName = self._get_variablename_from_dataset_schema(schemaName, metric_name)
+                    # variableName = self._get_variablename_from_dataset_schema(schemaName, metric_name)
+                    variableName = self._get_variablename_from_user_schema_mapping(schemaName.lower(), metric_name, schema_mapping)
                     mapping.append({"schemaName": schemaName, "variableName": variableName})
         return mapping
 
@@ -160,7 +187,7 @@ class Evaluation:
                         "model": "null",
                         "params": {
                             "model": {
-                                "value": "gpt-4o-mini"
+                                "value": ""
                             }
                         },
                         "mappings": "mappings"
@@ -178,7 +205,8 @@ class Evaluation:
         try:
             response = requests.get(
                 f'{self.base_url}/v1/llm/llm-metrics', 
-                headers=headers)
+                headers=headers,
+                timeout=self.timeout)
             response.raise_for_status()
             metrics_schema = [metric for metric in response.json()["data"]["metrics"]]
             return metrics_schema
@@ -208,14 +236,21 @@ class Evaluation:
                 #checking if provider is one of the allowed providers
                 if key.lower()=="provider" and value.lower() not in sub_providers:
                     raise ValueError("Enter a valid provider name. The following Provider names are supported: OpenAI, Azure, Gemini, Groq")
-                
-                base_json["metricSpec"]["config"]["params"][key] = {"value": value}
+    
+                if key.lower()=="threshold":
+                    if len(value)>1:
+                        raise ValueError("'threshold' can only take one argument gte/lte/eq")
+                    else:
+                        for key_thres, value_thres in value.items():
+                            base_json["metricSpec"]["config"]["params"][key] = {f"{key_thres}":value_thres}
+                else:
+                    base_json["metricSpec"]["config"]["params"][key] = {"value": value}
 
 
             # if metric["config"]["model"]:
             #     base_json["metricSpec"]["config"]["params"]["model"]["value"] = metric["config"]["model"]
             base_json["metricSpec"]["displayName"] = metric["column_name"]
-            mappings = self._get_mapping(metric["name"], metrics_schema_response)
+            mappings = self._get_mapping(metric["name"], metrics_schema_response, metric["schema_mapping"])
             base_json["metricSpec"]["config"]["mappings"] = mappings
             metricParams.append(base_json)
         metric_schema_mapping["metricParams"] = metricParams
@@ -228,12 +263,15 @@ class Evaluation:
         }
         try:
             response = requests.get(
-                f'{self.base_url}/v1/llm/filter?datasetId={str(self.dataset_id)}', 
-                headers=headers
-                )
+                f"{self.base_url}/v2/llm/dataset/{str(self.dataset_id)}?initialCols=0",
+                headers=headers,
+                timeout=self.timeout,
+            )
             response.raise_for_status()
-            executed_metric_response = response.json()["data"]["filter"]
-            executed_metric_list = [item["displayName"] for item in executed_metric_response]
+            dataset_columns = response.json()["data"]["datasetColumnsResponses"]
+            dataset_columns = [item["displayName"] for item in dataset_columns]
+            executed_metric_list = [data for data in dataset_columns if not data.startswith('_')]
+
             return executed_metric_list
         except requests.exceptions.HTTPError as http_err:
             logger.error(f"HTTP error occurred: {http_err}")
@@ -248,6 +286,13 @@ class Evaluation:
             return []
 
     def add_metrics(self, metrics):
+        #Handle required key if missing
+        required_keys = {"name", "config", "column_name", "schema_mapping"}
+        for metric in metrics:
+            missing_keys = required_keys - metric.keys()
+            if missing_keys:
+                raise ValueError(f"{missing_keys} required for each metric evaluation.")
+
         executed_metric_list = self._get_executed_metrics_list()
         metrics_name = self.list_metrics()
         user_metric_names = [metric["name"] for metric in metrics]
@@ -265,12 +310,12 @@ class Evaluation:
             'X-Project-Id': str(self.project_id),
         }
         metric_schema_mapping = self._update_base_json(metrics)
-        print(metric_schema_mapping)
         try:
             response = requests.post(
                 f'{self.base_url}/playground/metric-evaluation', 
                 headers=headers, 
-                json=metric_schema_mapping
+                json=metric_schema_mapping,
+                timeout=self.timeout
                 )
             if response.status_code == 400:
                 raise ValueError(response.json()["message"])
@@ -296,14 +341,20 @@ class Evaluation:
             "Authorization": f"Bearer {os.getenv('RAGAAI_CATALYST_TOKEN')}",
             'X-Project-Id': str(self.project_id),
         }
-        data = {"jobId": self.jobId}
         try:
-            response = requests.post(
+            response = requests.get(
                 f'{self.base_url}/job/status', 
                 headers=headers, 
-                json=data)
+                timeout=self.timeout)
             response.raise_for_status()
-            print(response.json()["data"]["status"])
+            if response.json()["success"]:
+                status_json = [item["status"] for item in response.json()["data"]["content"] if item["id"]==self.jobId][0]
+            if status_json == "Failed":
+                return print("Job failed. No results to fetch.")
+            elif status_json == "In Progress":
+                return print(f"Job in progress. Please wait while the job completes.\nVisit Job Status: {self.base_url.removesuffix('/api')}/projects/job-status?projectId={self.project_id} to track")
+            elif status_json == "Completed":
+                print(f"Job completed. Fetching results.\nVisit Job Status: {self.base_url.removesuffix('/api')}/projects/job-status?projectId={self.project_id} to check")
         except requests.exceptions.HTTPError as http_err:
             logger.error(f"HTTP error occurred: {http_err}")
         except requests.exceptions.ConnectionError as conn_err:
@@ -336,7 +387,8 @@ class Evaluation:
                 response = requests.post(
                     f'{self.base_url}/v1/llm/docs', 
                     headers=headers, 
-                    json=data)
+                    json=data,
+                    timeout=self.timeout)
                 response.raise_for_status()
                 return response.json()
             except requests.exceptions.HTTPError as http_err:
@@ -355,7 +407,7 @@ class Evaluation:
             try:
                 response = get_presignedUrl()
                 preSignedURL = response["data"]["preSignedURL"]
-                response = requests.get(preSignedURL)
+                response = requests.get(preSignedURL, timeout=self.timeout)
                 response.raise_for_status()
                 return response.text
             except requests.exceptions.HTTPError as http_err:

ragaai_catalyst/proxy_call.py

@@ -0,0 +1,134 @@
+import requests
+import json
+import subprocess
+import logging
+import traceback
+
+logger = logging.getLogger(__name__)
+
+def api_completion(model,messages, api_base='http://127.0.0.1:8000',
+                    api_key='',model_config=dict()):
+    whoami = get_username()
+    all_response = list()
+    job_id = model_config.get('job_id',-1)
+    converted_message = convert_input(messages,model,model_config)
+    payload = json.dumps(converted_message)
+    response = payload
+    headers = {
+        'Content-Type': 'application/json',
+        'Wd-PCA-Feature-Key':f'your_feature_key, $(whoami)'
+    }
+    try:
+        response = requests.request("POST", api_base, headers=headers, data=payload, verify=False)
+        if model_config.get('log_level','')=='debug':
+            logger.info(f'Model response Job ID {job_id} {response.text}')
+        if response.status_code!=200:
+            logger.error(f'Error in model response Job ID {job_id}:',str(response.text))
+            raise ValueError(str(response.text))
+    except Exception as e:
+        logger.error(f'Error in calling api Job ID {job_id}:',str(e))
+        raise ValueError(str(e))
+    try:
+        response = response.json()
+        if 'error' in response:
+            logger.error(f'Invalid response from API Job ID {job_id}:'+str(response))
+            raise ValueError(str(response.get('error')))
+        all_response.append(convert_output(response,job_id))
+    except ValueError as e1:
+        logger.error(f'Invalid json response from API Job ID {job_id}:'+response)
+        raise ValueError(str(e1))
+    except Exception as e1:
+        if model_config.get('log_level','')=='debug':
+            logger.info(f"Error trace Job ID: {job_id} {traceback.print_exc()}")
+        logger.error(f"Exception in parsing model response Job ID:{job_id} {str(e1)}")
+        logger.error(f"Model response Job ID: {job_id} {response.text}")
+        all_response.append(None)
+    return all_response
+
+def get_username():
+    result = subprocess.run(['whoami'], capture_output=True, text=True)
+    result = result.stdout
+    return result
+
+def convert_output(response,job_id):
+    try:
+        if response.get('prediction',{}).get('type','')=='generic-text-generation-v1':
+            return response['prediction']['output']
+        elif response.get('prediction',{}).get('type','')=='gcp-multimodal-v1':
+            full_response = ''
+            for chunk in response['prediction']['output']['chunks']:
+                candidate = chunk['candidates'][0]
+                if candidate['finishReason'] and candidate['finishReason'] not in ['STOP']:
+                    raise ValueError(candidate['finishReason'])
+                part = candidate['content']['parts'][0]
+                full_response += part['text']
+            return full_response
+        else:
+            raise ValueError('Invalid prediction type passed in config')
+    except ValueError as e1:
+        raise ValueError(str(e1))
+    except Exception as e:
+        logger.warning(f'Exception in formatting model response Job ID {job_id}:'+str(e))
+        return None
+
+
+def convert_input(prompt,model,model_config):
+    doc_input = {
+        "target": {
+            "provider": "echo",
+            "model": "echo"
+        },
+        "task": {
+            "type": "gcp-multimodal-v1",
+            "prediction_type": "gcp-multimodal-v1",
+            "input": {
+            "contents": [
+                {
+                "role": "user",
+                "parts": [
+                    {
+                    "text": "Give me a recipe for banana bread."
+                    }
+                ]
+                }
+            ],
+            "safetySettings": 
+                [
+                    {"category": "HARM_CATEGORY_HARASSMENT", "threshold": "BLOCK_NONE"},
+                    {"category": "HARM_CATEGORY_HATE_SPEECH", "threshold": "BLOCK_NONE"},
+                    {"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT", "threshold": "BLOCK_NONE"},
+                    {"category": "HARM_CATEGORY_DANGEROUS_CONTENT", "threshold": "BLOCK_NONE"},
+                ],
+            "generationConfig": {
+                "temperature": 0,
+                "maxOutputTokens": 8000,
+                "topK": 40,
+                "topP": 0.95,
+                "stopSequences": [],
+                "candidateCount": 1
+            }
+            }
+        }
+    }
+    if 'provider' not in model_config:
+        doc_input['target']['provider'] = 'gcp'
+    else:
+        doc_input['target']['provider'] = model_config['provider']
+    doc_input['task']['type'] = model_config.get('task_type','gcp-multimodal-v1')
+    doc_input['task']['prediction_type'] = model_config.get('prediction_type','generic-text-generation-v1')
+    if 'safetySettings' in model_config:
+        doc_input['task']['input']['safetySettings'] = model_config.get('safetySettings')
+    if 'generationConfig' in model_config:
+        doc_input['task']['input']['generationConfig'] = model_config.get('generationConfig')
+    doc_input['target']['model'] = model
+    if model_config.get('log_level','')=='debug':
+        logger.info(f"Using model configs Job ID {model_config.get('job_id',-1)}{doc_input}")
+    doc_input['task']['input']['contents'][0]['parts'] = [{"text":prompt[0]['content']}]
+    return doc_input
+
+
+
+if __name__=='__main__':
+    message_list = ["Hi How are you","I am good","How are you"]
+    response = batch_completion('gemini/gemini-1.5-flash',message_list,0,1,100,api_base='http://127.0.0.1:5000')
+    print(response)

ragaai_catalyst/synthetic_data_generation.py

@@ -0,0 +1,323 @@
+import os
+from groq import Groq
+import google.generativeai as genai
+import openai
+import PyPDF2
+import csv
+import markdown
+import pandas as pd
+import json
+from ragaai_catalyst import proxy_call
+import ast
+
+# dotenv.load_dotenv()
+
+class SyntheticDataGeneration:
+    """
+    A class for generating synthetic data using various AI models and processing different document types.
+    """
+
+    def __init__(self):
+        """
+        Initialize the SyntheticDataGeneration class with API clients for Groq, Gemini, and OpenAI.
+        """
+    def generate_qna(self, text, question_type="simple", n=5,model_config=dict(),api_key=None):
+        """
+        Generate questions based on the given text using the specified model and provider.
+
+        Args:
+            text (str): The input text to generate questions from.
+            question_type (str): The type of questions to generate ('simple', 'mcq', or 'complex').
+            model (str): The specific model to use for generation.
+            provider (str): The AI provider to use ('groq', 'gemini', or 'openai').
+            n (int): The number of question/answer pairs to generate.
+
+        Returns:
+            pandas.DataFrame: A DataFrame containing the generated questions and answers.
+
+        Raises:
+            ValueError: If an invalid provider is specified.
+        """
+        provider = model_config.get("provider")
+        model = model_config.get("model")
+        api_base = model_config.get("api_base")
+
+        system_message = self._get_system_message(question_type, n)
+        if provider == "groq":
+            if api_key is None and os.getenv("GROQ_API_KEY") is None:
+                raise ValueError("API key must be provided for Groq.")
+            self.groq_client = Groq(api_key=api_key or os.getenv("GROQ_API_KEY"))
+            return self._generate_groq(text, system_message, model)
+        elif provider == "gemini":
+            genai.configure(api_key=api_key or os.getenv("GEMINI_API_KEY"))
+            if api_base is None:
+                if api_key is None and os.getenv("GEMINI_API_KEY") is None:
+                    raise ValueError("API key must be provided for Gemini.")
+                genai.configure(api_key=api_key or os.getenv("GEMINI_API_KEY"))
+                return self._generate_gemini(text, system_message, model)
+            else:
+                messages=[
+                {'role': 'user', 'content': system_message+text}
+            ]
+                a= proxy_call.api_completion(messages=messages ,model=model ,api_base=api_base)
+                b= ast.literal_eval(a[0])
+                return pd.DataFrame(b)
+        elif provider == "openai":
+            if api_key is None and os.getenv("OPENAI_API_KEY") is None:
+                raise ValueError("API key must be provided for OpenAI.")
+            openai.api_key = api_key or os.getenv("OPENAI_API_KEY")
+            return self._generate_openai(text, system_message, model,api_key=api_key)
+        else:
+            raise ValueError("Invalid provider. Choose 'groq', 'gemini', or 'openai'.")
+
+    def _get_system_message(self, question_type, n):
+        """
+        Get the appropriate system message for the specified question type.
+
+        Args:
+            question_type (str): The type of questions to generate ('simple', 'mcq', or 'complex').
+            n (int): The number of question/answer pairs to generate.
+
+        Returns:
+            str: The system message for the AI model.
+
+        Raises:
+            ValueError: If an invalid question type is specified.
+        """
+        if question_type == 'simple':
+            return f'''Generate a set of {n} very simple questions answerable in a single phrase. 
+                Also return the answers for the generated questions.
+                Return the response in a list of object format. 
+                Each object in list should have Question and corresponding answer.
+                Do not return any extra strings. Return Generated text strictly in below format.  
+                [{{"Question":"question,"Answer":"answer"}}]
+            '''
+        elif question_type == 'mcq':
+            return f'''Generate a set of {n} questions with 4 probable answers from the given text. 
+                The options should not be longer than a phrase. There should be only 1 correct answer.
+                There should not be any ambiguity between correct and incorrect options.
+                Return the response in a list of object format. 
+                Each object in list should have Question and a list of options. 
+                Do not return any extra strings. Return Generated text strictly in below format. 
+                [{{"Question":"question","Options":[option1,option2,option3,option4]}}]
+            '''
+        elif question_type == 'complex':
+            return f'''Can you generate a set of {n} complex questions answerable in long form from the below texts.
+                Make sure the questions are important and provide new information to the user.
+                Return the response in a list of object format. Enclose any quotes in single quote. 
+                Do not use double quotes within questions or answers.
+                Each object in list should have Question and corresponding answer.
+                Do not return any extra strings. Return generated text strictly in below format.
+                [{{"Question":"question","Answer":"answers"}}]
+            '''
+        else:
+            raise ValueError("Invalid question type")
+
+    def _generate_groq(self, text, system_message, model):
+        """
+        Generate questions using the Groq API.
+
+        Args:
+            text (str): The input text to generate questions from.
+            system_message (str): The system message for the AI model.
+            model (str): The specific Groq model to use.
+
+        Returns:
+            pandas.DataFrame: A DataFrame containing the generated questions and answers.
+        """
+        response = self.groq_client.chat.completions.create(
+            model=model,
+            messages=[
+                {'role': 'system', 'content': system_message},
+                {'role': 'user', 'content': text}
+            ]
+        )
+        return self._parse_response(response, provider="groq")
+
+    def _generate_gemini(self, text, system_message, model):
+        """
+        Generate questions using the Gemini API.
+
+        Args:
+            text (str): The input text to generate questions from.
+            system_message (str): The system message for the AI model.
+            model (str): The specific Gemini model to use.
+
+        Returns:
+            pandas.DataFrame: A DataFrame containing the generated questions and answers.
+        """
+        model = genai.GenerativeModel(model)
+        response = model.generate_content([system_message, text])
+        return self._parse_response(response, provider="gemini")
+
+    def _generate_openai(self, text, system_message, model,api_key=None):
+        """
+        Generate questions using the OpenAI API.
+
+        Args:+
+            text (str): The input text to generate questions from.
+            system_message (str): The system message for the AI model.
+            model (str): The specific OpenAI model to use.
+
+        Returns:
+            pandas.DataFrame: A DataFrame containing the generated questions and answers.
+        """
+        client = openai.OpenAI(api_key=api_key)
+        response = client.chat.completions.create(
+            model=model,
+            messages=[
+                {"role": "system", "content": system_message},
+                {"role": "user", "content": text}
+            ]
+        )
+        return self._parse_response(response, provider="openai")
+
+    def _parse_response(self, response, provider):
+        """
+        Parse the response from the AI model and return it as a DataFrame.
+
+        Args:
+            response (str): The response from the AI model.
+            provider (str): The AI provider used ('groq', 'gemini', or 'openai').
+        Returns:
+            pandas.DataFrame: The parsed response as a DataFrame.
+        """
+        if provider == "openai":
+            data = response.choices[0].message.content
+        elif provider == "gemini":
+            data = response.candidates[0].content.parts[0].text
+        elif provider == "groq":
+            data = response.choices[0].message.content.replace('\n', '')
+            list_start_index = data.find('[')  # Find the index of the first '['
+            substring_data = data[list_start_index:] if list_start_index != -1 else data  # Slice from the list start
+            data = substring_data
+
+        else:
+            raise ValueError("Invalid provider. Choose 'groq', 'gemini', or 'openai'.")
+        try:
+            json_data = json.loads(data)
+            return pd.DataFrame(json_data)
+        except json.JSONDecodeError:
+            # If JSON parsing fails, return a DataFrame with a single column
+            return pd.DataFrame({'content': [data]})
+
+    def process_document(self, input_data):
+        """
+        Process the input document and extract its content.
+
+        Args:
+            input_data (str): Either a file path or a string of text.
+
+        Returns:
+            str: The extracted text content from the document.
+
+        Raises:
+            ValueError: If the input is neither a valid file path nor a string of text.
+        """
+        if isinstance(input_data, str):
+            if os.path.isfile(input_data):
+                # If input_data is a file path
+                _, file_extension = os.path.splitext(input_data)
+                if file_extension.lower() == '.pdf':
+                    return self._read_pdf(input_data)
+                elif file_extension.lower() == '.txt':
+                    return self._read_text(input_data)
+                elif file_extension.lower() == '.md':
+                    return self._read_markdown(input_data)
+                elif file_extension.lower() == '.csv':
+                    return self._read_csv(input_data)
+                else:
+                    raise ValueError(f"Unsupported file type: {file_extension}")
+            else:
+                # If input_data is a string of text
+                return input_data
+        else:
+            raise ValueError("Input must be either a file path or a string of text")
+
+    def _read_pdf(self, file_path):
+        """
+        Read and extract text from a PDF file.
+
+        Args:
+            file_path (str): The path to the PDF file.
+
+        Returns:
+            str: The extracted text content from the PDF.
+        """
+        text = ""
+        with open(file_path, 'rb') as file:
+            pdf_reader = PyPDF2.PdfReader(file)
+            for page in pdf_reader.pages:
+                text += page.extract_text()
+        return text
+
+    def _read_text(self, file_path):
+        """
+        Read the contents of a text file.
+
+        Args:
+            file_path (str): The path to the text file.
+
+        Returns:
+            str: The contents of the text file.
+        """
+        with open(file_path, 'r', encoding='utf-8') as file:
+            return file.read()
+
+    def _read_markdown(self, file_path):
+        """
+        Read and convert a Markdown file to HTML.
+
+        Args:
+            file_path (str): The path to the Markdown file.
+
+        Returns:
+            str: The HTML content converted from the Markdown file.
+        """
+        with open(file_path, 'r', encoding='utf-8') as file:
+            md_content = file.read()
+            html_content = markdown.markdown(md_content)
+            return html_content
+
+    def _read_csv(self, file_path):
+        """
+        Read and extract text from a CSV file.
+
+        Args:
+            file_path (str): The path to the CSV file.
+
+        Returns:
+            str: The extracted text content from the CSV, with each row joined and separated by newlines.
+        """
+        text = ""
+        with open(file_path, 'r', encoding='utf-8') as file:
+            csv_reader = csv.reader(file)
+            for row in csv_reader:
+                text += " ".join(row) + "\n"
+        return text
+
+    def get_supported_qna(self):
+        """
+        Get a list of supported question types.
+
+        Returns:
+            list: A list of supported question types.
+        """
+        return ['simple', 'mcq', 'complex']
+
+    def get_supported_providers(self):
+        """
+        Get a list of supported AI providers.
+
+        Returns:
+            list: A list of supported AI providers.
+        """
+        return ['gemini', 'openai']
+
+# Usage:
+# from synthetic_data_generation import SyntheticDataGeneration
+# synthetic_data_generation = SyntheticDataGeneration()
+# text = synthetic_data_generation.process_document(input_data=text_file)
+# result = synthetic_data_generation.generate_question(text)
+# supported_question_types = synthetic_data_generation.get_supported_question_types()
+# supported_providers = synthetic_data_generation.get_supported_providers()

{ragaai_catalyst-2.0.3.dist-info → ragaai_catalyst-2.0.5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ragaai_catalyst
-Version: 2.0.3
+Version: 2.0.5
 Summary: RAGA AI CATALYST
 Author-email: Kiran Scaria <kiran.scaria@raga.ai>, Kedar Gaikwad <kedar.gaikwad@raga.ai>, Dushyant Mahajan <dushyant.mahajan@raga.ai>, Siddhartha Kosti <siddhartha.kosti@raga.ai>, Ritika Goel <ritika.goel@raga.ai>, Vijay Chaurasia <vijay.chaurasia@raga.ai>
 Requires-Python: >=3.9
@@ -20,6 +20,10 @@ Requires-Dist: langchain-core>=0.2.11
 Requires-Dist: langchain>=0.2.11
 Requires-Dist: openai>=1.35.10
 Requires-Dist: pandas>=2.1.1
+Requires-Dist: groq>=0.11.0
+Requires-Dist: PyPDF2>=3.0.1
+Requires-Dist: google-generativeai>=0.8.2
+Requires-Dist: Markdown>=3.7
 Requires-Dist: tenacity==8.3.0
 Provides-Extra: dev
 Requires-Dist: pytest; extra == "dev"

{ragaai_catalyst-2.0.3.dist-info → ragaai_catalyst-2.0.5.dist-info}/RECORD

@@ -1,10 +1,12 @@
-ragaai_catalyst/__init__.py,sha256=kxR70lWGd_b5q6wn7wg8h_Oth5EohSqzMPjWMe1Za50,334
+ragaai_catalyst/__init__.py,sha256=T0-X4yfIAe26-tWx6kLwNkKIjaFoQL2aNLIRp5wBG5w,424
 ragaai_catalyst/_version.py,sha256=JKt9KaVNOMVeGs8ojO6LvIZr7ZkMzNN-gCcvryy4x8E,460
 ragaai_catalyst/dataset.py,sha256=XjI06Exs6-64pQPQlky4mtcUllNMCgKP-bnM_t9EWkY,10920
-ragaai_catalyst/evaluation.py,sha256=4cgCKPhFtnPdYsBcN8jTxTHcysycz-cQbau_MCiIUYg,16474
+ragaai_catalyst/evaluation.py,sha256=PR7rMkvZ4km26B24sSc60GPNS0JkrUMIYo5CPEqX2Qw,19315
 ragaai_catalyst/experiment.py,sha256=8KvqgJg5JVnt9ghhGDJvdb4mN7ETBX_E5gNxBT0Nsn8,19010
 ragaai_catalyst/prompt_manager.py,sha256=ZMIHrmsnPMq20YfeNxWXLtrxnJyMcxpeJ8Uya7S5dUA,16411
+ragaai_catalyst/proxy_call.py,sha256=nlMdJCSW73sfN0fMbCbtIk6W992Nac5FJvcfNd6UDJk,5497
 ragaai_catalyst/ragaai_catalyst.py,sha256=5Q1VCE7P33DtjaOtVGRUgBL8dpDL9kjisWGIkOyX4nE,17426
+ragaai_catalyst/synthetic_data_generation.py,sha256=STpZF-a1mYT3GR4CGdDvhBdctf2ciSLyvDANqJxnQp8,12989
 ragaai_catalyst/utils.py,sha256=TlhEFwLyRU690HvANbyoRycR3nQ67lxVUQoUOfTPYQ0,3772
 ragaai_catalyst/tracers/__init__.py,sha256=NppmJhD3sQ5R1q6teaZLS7rULj08Gb6JT8XiPRIe_B0,49
 ragaai_catalyst/tracers/tracer.py,sha256=eaGJdLEIjadHpbWBXBl5AhMa2vL97SVjik4U1L8gros,9591
@@ -17,7 +19,7 @@ ragaai_catalyst/tracers/instrumentators/llamaindex.py,sha256=SMrRlR4xM7k9HK43hak
 ragaai_catalyst/tracers/instrumentators/openai.py,sha256=14R4KW9wQCR1xysLfsP_nxS7cqXrTPoD8En4MBAaZUU,379
 ragaai_catalyst/tracers/utils/__init__.py,sha256=KeMaZtYaTojilpLv65qH08QmpYclfpacDA0U3wg6Ybw,64
 ragaai_catalyst/tracers/utils/utils.py,sha256=ViygfJ7vZ7U0CTSA1lbxVloHp4NSlmfDzBRNCJuMhis,2374
-ragaai_catalyst-2.0.3.dist-info/METADATA,sha256=gWehjEvbhpzWjNBDrYAD5RshBEOSL__-Dz-cejkOyvs,6497
-ragaai_catalyst-2.0.3.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
-ragaai_catalyst-2.0.3.dist-info/top_level.txt,sha256=HpgsdRgEJMk8nqrU6qdCYk3di7MJkDL0B19lkc7dLfM,16
-ragaai_catalyst-2.0.3.dist-info/RECORD,,
+ragaai_catalyst-2.0.5.dist-info/METADATA,sha256=tWppjo0sERHjjugIOAWdwD1p05HO6T6N_E1KYd9G9hY,6625
+ragaai_catalyst-2.0.5.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
+ragaai_catalyst-2.0.5.dist-info/top_level.txt,sha256=HpgsdRgEJMk8nqrU6qdCYk3di7MJkDL0B19lkc7dLfM,16
+ragaai_catalyst-2.0.5.dist-info/RECORD,,