Merge pull request #101 from patrickfleith/feature/moving-language-code-in-a-column

Feature/moving language code in a column

Author: patrickfleith

datafast/datasets.py

@@ -416,7 +416,7 @@ def generate(self, llms: list[LLMProvider]) -> "ClassificationDataset":
                                     label=label["name"],
                                     model_id=llm.model_id,
                                     label_source=LabelSource.SYNTHETIC,
-                                    metadata={"language": lang_code},
+                                    language=lang_code,
                                 )
                                 self.data_rows.append(row)
                                 new_rows.append(row)
@@ -510,8 +510,8 @@ def generate(self, llms: list[LLMProvider]) -> "RawDataset":
                                         text=example.text,
                                         text_source=TextSource.SYNTHETIC,
                                         model_id=llm.model_id,
+                                        language=lang_code,
                                         metadata={
-                                            "language": lang_code,
                                             "document_type": document_type,
                                             "topic": topic,
                                         },
@@ -686,8 +686,8 @@ def generate(self, llms: list[LLMProvider]) -> "UltrachatDataset":
                                         opening_question=messages[0]["content"],
                                         messages=messages,
                                         model_id=llm.model_id,
+                                        language=lang_code,
                                         metadata={
-                                            "language": lang_code,
                                             "domain": self.config.domain,
                                             "topic": topic,
                                             "subtopic": subtopic,
@@ -915,8 +915,8 @@ def generate(self, llms: list[LLMProvider]) -> "MCQDataset":
                                                 incorrect_answer_3=incorrect_answers[2],
                                                 model_id=llm.model_id,
                                                 mcq_source=MCQSource.SYNTHETIC,
+                                                language=lang_code,
                                                 metadata={
-                                                    "language": lang_code,
                                                     "source_dataset": self._get_source_dataset_name(),
                                                 },
                                             )
@@ -1092,8 +1092,8 @@ def generate(self,
                         "preference_source": PreferenceSource.SYNTHETIC,
                         "chosen_model_id": chosen_model_id,
                         "rejected_model_id": rejected_model_id,
+                        "language": lang_code,
                         "metadata": {
-                            "language": lang_code,
                             "instruction_model": question_gen_llm.model_id,
                         }
                     }

datafast/examples/quickstart_example.py

@@ -25,7 +25,7 @@
         ],
     expansion=PromptExpansionConfig(
         placeholders={
-            "context": ["hike review", "speedboat tour review", "outdoor climbing experience"],
+            "context": ["hike review", "speedboat tour review"],
         },
         combinatorial=True
     )
@@ -34,7 +34,7 @@
 from datafast.llms import OpenAIProvider, AnthropicProvider, GeminiProvider, OllamaProvider
 
 providers = [
-    OpenAIProvider(model_id="gpt-4.1-nano"),
+    OpenAIProvider(model_id="gpt-5-nano-2025-08-07"),
     # AnthropicProvider(model_id="claude-3-5-haiku-latest"),
     # GeminiProvider(model_id="gemini-2.0-flash"),
     # OllamaProvider(model_id="gemma3:12b")
@@ -47,9 +47,9 @@
 dataset.generate(providers)
 
 # Optional: Push to Hugging Face Hub
-USERNAME = "username"  # <--- Your hugging face username
-DATASET_NAME = "dataset_name"  # <--- Your hugging face dataset name
+USERNAME = "patrickfleith"  # <--- Your hugging face username
+DATASET_NAME = "datafast_quickstart_no_train_test_split"  # <--- Your hugging face dataset name
 dataset.push_to_hub(
     repo_id=f"{USERNAME}/{DATASET_NAME}",
-    train_size=0.6
+    # train_size=0.6
 )

datafast/examples/show_dataset_examples.py

@@ -38,7 +38,7 @@
     text="The trail is blocked by a fallen tree.",
     label="trail_obstruction",
     model_id="gpt-4.1-nano",
-    metadata={"language": "en"},
+    language="en",
 )
 classification_dataset = ClassificationDataset(
     ClassificationDatasetConfig(classes=[{"name": "trail_obstruction", "description": "Obstruction on the trail."}])
@@ -47,7 +47,7 @@
     text="The trail is well maintained and easy to follow.",
     label="positive_conditions",
     model_id="claude-3-5-haiku-latest",
-    metadata={"language": "en"},
+    language="en",
 )
 classification_dataset.data_rows = [classification_row, classification_row2]
 
@@ -60,7 +60,7 @@
     incorrect_answer_2="Berlin",
     incorrect_answer_3="Rome",
     model_id="gemini-2.0-flash",
-    metadata={"language": "en"},
+    language="en",
 )
 mcq_config = MCQDatasetConfig(
     text_column="source_document",
@@ -75,7 +75,7 @@
     incorrect_answer_2="Yangtze River",
     incorrect_answer_3="Mississippi River",
     model_id="gpt-4.1-nano",
-    metadata={"language": "en"},
+    language="en",
 )
 mcq_dataset.data_rows = [mcq_row, mcq_row2]
 
@@ -91,7 +91,7 @@
     rejected_response_score=3,
     chosen_response_assessment="Accurate and detailed.",
     rejected_response_assessment="Too generic.",
-    metadata={"language": "en"},
+    language="en",
 )
 preference_dataset = PreferenceDataset(PreferenceDatasetConfig(input_documents=["Describe a recent Mars mission."]))
 preference_row2 = PreferenceRow(
@@ -105,7 +105,7 @@
     rejected_response_score=2,
     chosen_response_assessment="Factually correct and detailed.",
     rejected_response_assessment="Incorrect mission.",
-    metadata={"language": "en"},
+    language="en",
 )
 preference_dataset.data_rows = [preference_row, preference_row2]
 
@@ -134,20 +134,21 @@
     persona="space policy expert",
     messages=[{"role": "user", "content": "What are current efforts to clean up space debris?"}, {"role": "assistant", "content": "There are several ongoing projects, such as RemoveDEBRIS and ClearSpace-1."}],
     model_id="gpt-4.1-nano",
-    metadata={"language": "en"}
+    language="en"
 )
 ultrachat_row2 = ChatRow(
     opening_question="What is the importance of the Moon missions?",
     persona="lunar geologist",
     messages=[{"role": "user", "content": "Why do we keep returning to the Moon?"}, {"role": "assistant", "content": "The Moon offers scientific insights and is a stepping stone for Mars exploration."}],
     model_id="gemini-2.0-flash",
-    metadata={"language": "en"}
+    language="en"
 )
 ultrachat_config = UltrachatDatasetConfig()
 ultrachat_dataset = UltrachatDataset(ultrachat_config)
 ultrachat_dataset.data_rows = [ultrachat_row1, ultrachat_row2]
 
 if __name__ == "__main__":
+    # pass
     # print("Showing ClassificationDataset example...")
     # inspect_classification_dataset(classification_dataset)
     # print("Showing MCQDataset example...")
@@ -156,5 +157,5 @@
     # inspect_preference_dataset(preference_dataset)
     # print("Showing RawDataset example...")
     # inspect_raw_dataset(raw_dataset)
-    # print("Showing UltrachatDataset example...")
-    # inspect_ultrachat_dataset(ultrachat_dataset)
+    print("Showing UltrachatDataset example...")
+    inspect_ultrachat_dataset(ultrachat_dataset)

datafast/inspectors.py

@@ -353,8 +353,8 @@ def format_conversation(self, conversation: List[Dict]) -> str:
     def show_example(self, idx: int) -> Tuple[str, str, str, Dict]:
         """Extract data from the example to display."""
         row = self.get_example(idx)
-        conversation = row.get("conversation", [])
-        formatted_convo = self.format_conversation(conversation)
+        messages = row.get("messages", [])
+        formatted_convo = self.format_conversation(messages)
         return (
             self.get_index_label(idx),
             formatted_convo,

datafast/schema/data_rows.py

@@ -1,6 +1,6 @@
 from pydantic import BaseModel, Field
 from uuid import UUID, uuid4
-from typing import Union, Optional
+from typing import Union
 from enum import Enum
 
 
@@ -33,7 +33,8 @@ class TextRow(BaseModel):
 
     text: str
     text_source: TextSource = TextSource.SYNTHETIC
-    model_id: Optional[str] = None
+    model_id: str | None = None
+    language: str | None = None
     uuid: UUID = Field(default_factory=uuid4)
     metadata: dict[str, str] = Field(default_factory=dict)
 
@@ -43,7 +44,8 @@ class ChatRow(BaseModel):
 
     opening_question: str
     messages: list[dict[str, str]]
-    model_id: Optional[str] = None
+    model_id: str | None = None
+    language: str | None = None
     uuid: UUID = Field(default_factory=uuid4)
     metadata: dict[str, str] = Field(default_factory=dict)
     persona: str
@@ -52,9 +54,10 @@ class ChatRow(BaseModel):
 class TextClassificationRow(BaseModel):
     text: str
     label: LabelType  # Must be either str, list[str], or list[int]
-    model_id: Optional[str] = None
+    model_id: str | None = None
     label_source: LabelSource = LabelSource.SYNTHETIC
-    confidence_scores: Optional[dict[str, float]] = Field(default_factory=dict)
+    confidence_scores: dict[str, float] | None = Field(default_factory=dict)
+    language: str | None = None
 
     # System and metadata fields
     uuid: UUID = Field(default_factory=uuid4)
@@ -68,8 +71,9 @@ class MCQRow(BaseModel):
     incorrect_answer_1: str
     incorrect_answer_2: str
     incorrect_answer_3: str
-    model_id: Optional[str] = None
+    model_id: str | None = None
     mcq_source: MCQSource = MCQSource.SYNTHETIC
+    language: str | None = None
     uuid: UUID = Field(default_factory=uuid4)
     metadata: dict[str, str] = Field(default_factory=dict)
 
@@ -89,14 +93,15 @@ class PreferenceRow(BaseModel):
     chosen_response: str
     rejected_response: str
     preference_source: PreferenceSource = PreferenceSource.SYNTHETIC
-    chosen_model_id: Optional[str] = None
-    rejected_model_id: Optional[str] = None
+    chosen_model_id: str | None = None
+    rejected_model_id: str | None = None
+    language: str | None = None
 
     # Optional judge-related fields
-    chosen_response_score: Optional[int] = None
-    rejected_response_score: Optional[int] = None
-    chosen_response_assessment: Optional[str] = None
-    rejected_response_assessment: Optional[str] = None
+    chosen_response_score: int | None = None
+    rejected_response_score: int | None = None
+    chosen_response_assessment: str | None = None
+    rejected_response_assessment: str | None = None
 
     uuid: UUID = Field(default_factory=uuid4)
     metadata: dict[str, str] = Field(default_factory=dict)