chunkr-ai 0.0.45__py3-none-any.whl → 0.0.46__py3-none-any.whl

chunkr_ai/api/configuration.py

@@ -1,7 +1,7 @@
 from pydantic import BaseModel, Field, ConfigDict
 from enum import Enum
 from typing import Any, List, Optional, Union
-from pydantic import field_validator
+from pydantic import field_validator, field_serializer
 
 class GenerationStrategy(str, Enum):
     LLM = "LLM"
@@ -65,11 +65,7 @@ class TokenizerType(BaseModel):
             return f"string:{self.string_value}"
         return ""
     
-    model_config = ConfigDict(
-        json_encoders={
-            'TokenizerType': lambda v: v.model_dump()
-        }
-    )
+    model_config = ConfigDict()
     
     def model_dump(self, **kwargs):
         if self.enum_value is not None:
@@ -85,10 +81,13 @@ class ChunkProcessing(BaseModel):
 
     model_config = ConfigDict(
         arbitrary_types_allowed=True,
-        json_encoders={
-            TokenizerType: lambda v: v.model_dump()
-        }
     )
+    
+    @field_serializer('tokenizer')
+    def serialize_tokenizer(self, tokenizer: Optional[TokenizerType], _info):
+        if tokenizer is None:
+            return None
+        return tokenizer.model_dump()
 
     @field_validator('tokenizer', mode='before')
     def validate_tokenizer(cls, v):
@@ -130,6 +129,95 @@ class ErrorHandlingStrategy(str, Enum):
     FAIL = "Fail"
     CONTINUE = "Continue"
 
+class FallbackStrategy(BaseModel):
+    type: str
+    model_id: Optional[str] = None
+    
+    @classmethod
+    def none(cls) -> "FallbackStrategy":
+        return cls(type="None")
+    
+    @classmethod
+    def default(cls) -> "FallbackStrategy":
+        return cls(type="Default")
+    
+    @classmethod
+    def model(cls, model_id: str) -> "FallbackStrategy":
+        return cls(type="Model", model_id=model_id)
+    
+    def __str__(self) -> str:
+        if self.type == "Model":
+            return f"Model({self.model_id})"
+        return self.type
+    
+    def model_dump(self, **kwargs):
+        if self.type == "Model":
+            return {"Model": self.model_id}
+        return self.type
+    
+    @field_validator('type')
+    def validate_type(cls, v):
+        if v not in ["None", "Default", "Model"]:
+            raise ValueError(f"Invalid fallback strategy: {v}")
+        return v
+    
+    model_config = ConfigDict()
+    
+    @classmethod
+    def model_validate(cls, obj):
+        # Handle string values like "None" or "Default"
+        if isinstance(obj, str):
+            if obj in ["None", "Default"]:
+                return cls(type=obj)
+            # Try to parse as Enum value if it's not a direct match
+            try:
+                return cls(type=obj)
+            except ValueError:
+                pass  # Let it fall through to normal validation
+                
+        # Handle dictionary format like {"Model": "model-id"}
+        elif isinstance(obj, dict) and len(obj) == 1:
+            if "Model" in obj:
+                return cls(type="Model", model_id=obj["Model"])
+        
+        # Fall back to normal validation
+        return super().model_validate(obj)
+
+class LlmProcessing(BaseModel):
+    model_id: Optional[str] = None
+    fallback_strategy: FallbackStrategy = Field(default_factory=FallbackStrategy.default)
+    max_completion_tokens: Optional[int] = None
+    temperature: float = 0.0
+    
+    model_config = ConfigDict()
+    
+    @field_serializer('fallback_strategy')
+    def serialize_fallback_strategy(self, fallback_strategy: FallbackStrategy, _info):
+        return fallback_strategy.model_dump()
+
+    @field_validator('fallback_strategy', mode='before')
+    def validate_fallback_strategy(cls, v):
+        if isinstance(v, str):
+            if v == "None":
+                return FallbackStrategy.none()
+            elif v == "Default":
+                return FallbackStrategy.default()
+            # Try to parse as a model ID if it's not None or Default
+            try:
+                return FallbackStrategy.model(v)
+            except ValueError:
+                pass  # Let it fall through to normal validation
+        # Handle dictionary format like {"Model": "model-id"}
+        elif isinstance(v, dict) and len(v) == 1:
+            if "Model" in v:
+                return FallbackStrategy.model(v["Model"])
+            elif "None" in v or v.get("None") is None:
+                return FallbackStrategy.none()
+            elif "Default" in v or v.get("Default") is None:
+                return FallbackStrategy.default()
+                
+        return v
+
 class BoundingBox(BaseModel):
     left: float
     top: float
@@ -199,6 +287,7 @@ class Configuration(BaseModel):
     segment_processing: Optional[SegmentProcessing] = None
     segmentation_strategy: Optional[SegmentationStrategy] = None
     pipeline: Optional[Pipeline] = None
+    llm_processing: Optional[LlmProcessing] = None
     
 class OutputConfiguration(Configuration):
     input_file_url: Optional[str] = None

chunkr_ai/models.py

@@ -6,8 +6,10 @@ from .api.configuration import (
     CroppingStrategy,
     EmbedSource,
     ErrorHandlingStrategy,
+    FallbackStrategy,
     GenerationStrategy,
     GenerationConfig,
+    LlmProcessing,
     Model,
     OCRResult,
     OcrStrategy,
@@ -31,8 +33,10 @@ __all__ = [
     "CroppingStrategy",
     "EmbedSource",
     "ErrorHandlingStrategy",
+    "FallbackStrategy",
     "GenerationConfig",
     "GenerationStrategy",
+    "LlmProcessing",
     "Model",
     "OCRResult",
     "OcrStrategy",

{chunkr_ai-0.0.45.dist-info → chunkr_ai-0.0.46.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: chunkr-ai
-Version: 0.0.45
+Version: 0.0.46
 Summary: Python client for Chunkr: open source document intelligence
 Author-email: Ishaan Kapoor <ishaan@lumina.sh>
 License: MIT License

{chunkr_ai-0.0.45.dist-info → chunkr_ai-0.0.46.dist-info}/RECORD

@@ -1,16 +1,16 @@
 chunkr_ai/__init__.py,sha256=6KpYv2lmD6S5z2kc9pqwuLP5VDHmOuu2qDZArUIhb1s,53
-chunkr_ai/models.py,sha256=m3_3sUm4zm2LTqmxFWJtAGMqJbksmzCF2X2wGCAgSLY,920
+chunkr_ai/models.py,sha256=L0L9CjY8SgSh9_Fzvo_nJXqKf_2urZHngMWtBVlAQAo,1006
 chunkr_ai/api/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 chunkr_ai/api/auth.py,sha256=hlv0GiUmlsbFO1wLL9sslqOnsBSoBqkL_6Mk2SDvxgE,413
 chunkr_ai/api/chunkr.py,sha256=BzwcKNCuLfVR-HzgY8tKStsW4pIDVVjBgnEqPLyUUMM,3292
 chunkr_ai/api/chunkr_base.py,sha256=FDl0Ew8eOY4hur5FFqPENZiq9YQy0G3XWEqcKPeCO-U,6130
-chunkr_ai/api/configuration.py,sha256=KHOt1MiRvDbIvAdFK8I5j5uA4O561WeCBd1J1megQ7Y,6671
+chunkr_ai/api/configuration.py,sha256=aCYi_NjuTDynDc6g_N94jVGTb8SQQaUQ4LM8_a5v29g,9882
 chunkr_ai/api/decorators.py,sha256=VJX4qGBIL00K2zY8bh5KAMWv7SltJ38TvPJH06FnFss,4415
 chunkr_ai/api/misc.py,sha256=QN-2YWQ8e3VvvK63Ua-e8jsx6gxVxkO88Z96yWOofu0,3653
 chunkr_ai/api/protocol.py,sha256=LjPrYSq52m1afIlAo0yVGXlGZxPRh8J6g7S4PAit3Zo,388
 chunkr_ai/api/task_response.py,sha256=6kk9g2f7OZB3PAsmp4Or5A42r1dXTAzWAHEIVtLQ9sA,6545
-chunkr_ai-0.0.45.dist-info/licenses/LICENSE,sha256=w3R12yNDyZpMiy2lxy_hvNbsldC75ww79sF0u11rkho,1069
-chunkr_ai-0.0.45.dist-info/METADATA,sha256=owMljpIYmTevXkW9V-Nly0dBILzmyN8MfKeXpXctcVU,7053
-chunkr_ai-0.0.45.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
-chunkr_ai-0.0.45.dist-info/top_level.txt,sha256=0IZY7PZIiS8bw5r4NUQRUQ-ATi-L_3vLQVq3ZLouOW8,10
-chunkr_ai-0.0.45.dist-info/RECORD,,
+chunkr_ai-0.0.46.dist-info/licenses/LICENSE,sha256=w3R12yNDyZpMiy2lxy_hvNbsldC75ww79sF0u11rkho,1069
+chunkr_ai-0.0.46.dist-info/METADATA,sha256=Zjo2enHVCP5x0QqMTcS0k20nAWKogUoL88LZEVFoMZ8,7053
+chunkr_ai-0.0.46.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
+chunkr_ai-0.0.46.dist-info/top_level.txt,sha256=0IZY7PZIiS8bw5r4NUQRUQ-ATi-L_3vLQVq3ZLouOW8,10
+chunkr_ai-0.0.46.dist-info/RECORD,,