Bug/125 invalid json qa entries

README.md

@@ -46,7 +46,7 @@ pip install datafast
 
 ### 1. Environment Setup
 
-Make sure you have created a `secrets.env` file with your API keys.
+Make sure you have created a `.env` file with your API keys.
 HF token is needed if you want to push the dataset to your HF hub.
 Other keys depends on which LLM providers you use.
 ```
@@ -64,7 +64,7 @@ from datafast.llms import OpenAIProvider, AnthropicProvider, GeminiProvider
 from dotenv import load_dotenv
 
 # Load environment variables
-load_dotenv("secrets.env") # <--- your API keys
+load_dotenv() # <--- your API keys
 ```
 
 ### 3. Configure Dataset

datafast/__init__.py

@@ -1,6 +1,7 @@
 """Datafast - A Python package for synthetic text dataset generation"""
 
 import importlib.metadata
+from datafast.logger_config import configure_logger
 
 try:
     __version__ = importlib.metadata.version("datafast")
@@ -11,3 +12,5 @@
 def get_version():
     """Return the current version of the datafast package."""
     return __version__
+
+__all__ = ["configure_logger", "get_version"]

datafast/examples/classification_trail_conditions_example.py

@@ -1,10 +1,14 @@
 from datafast.datasets import ClassificationDataset
 from datafast.schema.config import ClassificationDatasetConfig, PromptExpansionConfig
 from datafast.llms import OpenAIProvider, AnthropicProvider
+from datafast.logger_config import configure_logger
 from dotenv import load_dotenv
 
 # Load API keys
-load_dotenv("secrets.env")
+load_dotenv()
+
+# Configure logger
+configure_logger()
 
 # Configure dataset
 config = ClassificationDatasetConfig(

datafast/examples/generic_pipeline_example.py

@@ -7,6 +7,8 @@
 from datafast.schema.config import GenericPipelineDatasetConfig
 from datafast.datasets import GenericPipelineDataset
 from datafast.llms import OpenAIProvider, GeminiProvider, OllamaProvider
+from datafast.logger_config import configure_logger
+from dotenv import load_dotenv
 
 
 PROMPT_TEMPLATE = """I will give you a persona.
@@ -80,7 +82,6 @@ def main():
 
 
 if __name__ == "__main__":
-    from dotenv import load_dotenv
-
-    load_dotenv("secrets.env")
+    load_dotenv()
+    configure_logger()
     main()

datafast/examples/generic_pipeline_response_format_example.py

@@ -2,6 +2,10 @@
 
 from datafast.schema.config import GenericPipelineDatasetConfig
 from datafast.utils import create_response_model
+from datafast.logger_config import configure_logger
+
+# Configure logger
+configure_logger()
 
 # Test with multiple columns and num_samples_per_prompt = 3
 config = GenericPipelineDatasetConfig(

datafast/examples/generic_pipeline_row_model_example.py

@@ -2,6 +2,10 @@
 
 from datafast.schema.config import GenericPipelineDatasetConfig
 from datafast.utils import create_generic_pipeline_row_model
+from datafast.logger_config import configure_logger
+
+# Configure logger
+configure_logger()
 
 # Test with multiple input, forward, and output columns
 config = GenericPipelineDatasetConfig(

datafast/examples/inspect_dataset_example.py

@@ -5,15 +5,19 @@
     python -m datafast.examples.inspect_dataset_example
 
 Requires:
-    - OpenAI API key in secrets.env or environment
+    - OpenAI API key in .env or environment
     - gradio package (pip install gradio)
 """
 from datafast.datasets import ClassificationDataset
 from datafast.schema.config import ClassificationDatasetConfig, PromptExpansionConfig
+from datafast.logger_config import configure_logger
 from dotenv import load_dotenv
 
-# Load API keys from environment or secrets.env
-load_dotenv("secrets.env")
+# Load API keys from environment or .env
+load_dotenv()
+
+# Configure logger
+configure_logger()
 
 # Configure the dataset generation
 config = ClassificationDatasetConfig(

datafast/examples/keywords_extraction_example.py

@@ -7,6 +7,12 @@
 from datafast.schema.config import GenericPipelineDatasetConfig
 from datafast.datasets import GenericPipelineDataset
 from datafast.llms import OpenAIProvider, GeminiProvider, OllamaProvider, OpenRouterProvider
+from datafast.logger_config import configure_logger
+from dotenv import load_dotenv
+
+# Load environment variables and configure logger
+load_dotenv()
+configure_logger()
 
 PROMPT_TEMPLATE = """I will give you a tweet.
 Generate a comma separated list of 3 keywords for the tweet. Avoid multi-word keywords.
@@ -63,7 +69,4 @@ def main():
 
 
 if __name__ == "__main__":
-    from dotenv import load_dotenv
-
-    load_dotenv("secrets.env")
     main()

datafast/examples/mcq_contextual_example.py

@@ -10,11 +10,15 @@
 import pandas as pd
 from dotenv import load_dotenv
 
-load_dotenv("secrets.env")
+load_dotenv()
 
 from datafast.schema.config import MCQDatasetConfig
 from datafast.datasets import MCQDataset
 from datafast.llms import OpenAIProvider
+from datafast.logger_config import configure_logger
+
+# Configure logger
+configure_logger()
 
 def main():
     # 1. Create a temporary filtered version of the dataset

datafast/examples/mcq_example.py

@@ -5,7 +5,13 @@
 import os
 from datafast.schema.config import MCQDatasetConfig
 from datafast.datasets import MCQDataset
-from datafast.llms import OpenAIProvider, AnthropicProvider, GeminiProvider
+from datafast.llms import OpenAIProvider, OpenRouterProvider
+from datafast.logger_config import configure_logger
+from dotenv import load_dotenv
+
+# Load environment variables and configure logger
+load_dotenv()
+configure_logger()
 
 
 def main():
@@ -16,8 +22,8 @@ def main():
         # local_file_path="datafast/examples/data/mcq/sample.txt",
         #local_file_path="datafast/examples/data/mcq/sample.jsonl",
         text_column="text",    # Column containing the text to generate questions from
-        sample_count=2,          # Process only 3 samples for testing
-        num_samples_per_prompt=2,# Generate 2 questions per document
+        sample_count=20,          # Process only 3 samples for testing
+        num_samples_per_prompt=3,# Generate 2 questions per document
         min_document_length=100, # Skip documents shorter than 100 chars
         max_document_length=20000,# Skip documents longer than 20000 chars
         output_file="mcq_test_dataset.jsonl",
@@ -26,13 +32,12 @@ def main():
     # 2. Initialize LLM providers
     providers = [
         OpenAIProvider(model_id="gpt-5-mini-2025-08-07"),
-        # AnthropicProvider(model_id="claude-haiku-4-5-20251001"),
-        # GeminiProvider(model_id="gemini-2.0-flash"),
+        OpenRouterProvider(model_id="qwen/qwen3-next-80b-a3b-instruct"),
     ]
 
     # 3. Generate the dataset
     dataset = MCQDataset(config)
-    num_expected_rows = dataset.get_num_expected_rows(providers, source_data_num_rows=2)
+    num_expected_rows = dataset.get_num_expected_rows(providers, source_data_num_rows=20)
     print(f"\nExpected number of rows: {num_expected_rows}")
     dataset.generate(providers)
 
@@ -53,7 +58,4 @@ def main():
 
 
 if __name__ == "__main__":
-    from dotenv import load_dotenv
-
-    load_dotenv()
     main()

datafast/examples/preference_dataset_example.py

@@ -78,5 +78,5 @@ def main():
 
 if __name__ == "__main__":
     from dotenv import load_dotenv
-    load_dotenv("secrets.env")
+    load_dotenv()
     main()

datafast/examples/quickstart_example.py

@@ -1,7 +1,10 @@
 from datafast.datasets import ClassificationDataset
 from datafast.schema.config import ClassificationDatasetConfig, PromptExpansionConfig
 from dotenv import load_dotenv
-load_dotenv("secrets.env")
+from datafast.logger_config import configure_logger
+load_dotenv()
+
+configure_logger()
 
 config = ClassificationDatasetConfig(
     classes=[

datafast/examples/raw_text_space_engineering_example.py

@@ -1,23 +1,26 @@
 from datafast.datasets import RawDataset
 from datafast.schema.config import RawDatasetConfig, PromptExpansionConfig
 from datafast.llms import OpenAIProvider, AnthropicProvider
+from datafast.logger_config import configure_logger
+
+configure_logger()
 
 
 def main():
     # 1. Configure the dataset generation
     config = RawDatasetConfig(
         document_types=[
-            "space engineering textbook", 
-            "spacecraft design justification document", 
+            # "space engineering textbook", 
+            # "spacecraft design justification document", 
             "personal blog of a space engineer"
         ],
         topics=[
-            "Microgravity",
-            "Vacuum",
-            "Heavy Ions",
-            "Thermal Extremes",
-            "Atomic Oxygen",
-            "Debris Impact",
+            # "Microgravity",
+            # "Vacuum",
+            # "Heavy Ions",
+            # "Thermal Extremes",
+            # "Atomic Oxygen",
+            # "Debris Impact",
             "Electrostatic Charging",
             "Propellant Boil-off",
             # ... You can pour hundreds of topics here. 8 is enough for this example
@@ -67,5 +70,5 @@ def main():
 if __name__ == "__main__":
     from dotenv import load_dotenv
 
-    load_dotenv("secrets.env")
+    load_dotenv()
     main()

datafast/examples/ultrachat_materials_science.py

@@ -53,5 +53,5 @@ def main():
 
 if __name__ == "__main__":
     from dotenv import load_dotenv
-    load_dotenv("secrets.env")
+    load_dotenv()
     main()

datafast/llms.py

@@ -10,6 +10,7 @@
 import time
 import traceback
 import warnings
+from loguru import logger
 
 # Pydantic
 from pydantic import BaseModel
@@ -65,6 +66,9 @@ def __init__(
 
         # Configure environment with API key if needed
         self._configure_env()
+
+        # Log successful initialization
+        logger.info(f"Initialized {self.provider_name} | Model: {self.model_id}")
 
     @property
     @abstractmethod
@@ -82,6 +86,9 @@ def _get_api_key(self) -> str:
         """Get API key from environment variables."""
         api_key = os.getenv(self.env_key_name)
         if not api_key:
+            logger.error(
+                f"Missing API key | Set {self.env_key_name} environment variable"
+            )
             raise ValueError(
                 f"{self.env_key_name} environment variable not set. "
                 f"Please set it or provide an API key when initializing the provider."
@@ -112,7 +119,7 @@ def _respect_rate_limit(self) -> None:
         # Add a 1s margin to avoid accidental rate limit exceedance
         sleep_time = 61 - (current - earliest)
         if sleep_time > 0:
-            print("Waiting for rate limit...")
+            logger.warning(f"Rate limit reached | Waiting {sleep_time:.1f}s")
             time.sleep(sleep_time)
 
     @staticmethod
@@ -264,8 +271,23 @@ def generate(
                 if response_format is not None:
                     # Strip code fences before validation
                     content = self._strip_code_fences(content)
-                    results.append(
-                        response_format.model_validate_json(content))
+                    try:
+                        results.append(
+                            response_format.model_validate_json(content))
+                    except Exception as validation_error:
+                        # Show the content that failed to parse for debugging
+                        content_preview = content[:200] + "..." if len(content) > 200 else content
+                        logger.warning(
+                            f"JSON parsing failed, skipping response | "
+                            f"Model: {self.model_id} | "
+                            f"Format: {response_format.__name__} | "
+                            f"Content preview: {content_preview}"
+                        )
+                        raise ValueError(
+                            f"Failed to parse JSON response into {response_format.__name__}.\n"
+                            f"Validation error: {validation_error}\n"
+                            f"Content received (first 200 chars):\n{content_preview}"
+                        ) from validation_error
                 else:
                     # Strip leading/trailing whitespace for text responses
                     results.append(content.strip() if content else content)
@@ -277,6 +299,10 @@ def generate(
 
         except Exception as e:
             error_trace = traceback.format_exc()
+            logger.error(
+                f"Generation failed | Provider: {self.provider_name} | "
+                f"Model: {self.model_id} | Error: {str(e)}"
+            )
             raise RuntimeError(
                 f"Error generating batch response with {self.provider_name}:\n{error_trace}"
             )
@@ -471,7 +497,22 @@ def generate(
                 if response_format is not None:
                     # Strip code fences before validation
                     content = self._strip_code_fences(content)
-                    results.append(response_format.model_validate_json(content))
+                    try:
+                        results.append(response_format.model_validate_json(content))
+                    except Exception as validation_error:
+                        # Show the content that failed to parse for debugging
+                        content_preview = content[:200] + "..." if len(content) > 200 else content
+                        logger.warning(
+                            f"JSON parsing failed, skipping response | "
+                            f"Model: {self.model_id} | "
+                            f"Format: {response_format.__name__} | "
+                            f"Content preview: {content_preview}"
+                        )
+                        raise ValueError(
+                            f"Failed to parse JSON response into {response_format.__name__}.\n"
+                            f"Validation error: {validation_error}\n"
+                            f"Content received (first 200 chars):\n{content_preview}"
+                        ) from validation_error
                 else:
                     # Strip leading/trailing whitespace for text responses
                     results.append(content.strip() if content else content)
@@ -483,6 +524,10 @@ def generate(
 
         except Exception as e:
             error_trace = traceback.format_exc()
+            logger.error(
+                f"Generation failed | Provider: {self.provider_name} | "
+                f"Model: {self.model_id} | Error: {str(e)}"
+            )
             raise RuntimeError(
                 f"Error generating response with {self.provider_name}:\n{error_trace}"
             )