Removing marshmallow-jsonschema

crates/indexify_internal_api/src/lib.rs

@@ -379,6 +379,7 @@ impl From<ContentMetadata> for indexify_coordinator::ContentMetadata {
 }
 
 // FIXME - Make this visible to only tests
+#[cfg(test)]
 impl Default for ContentMetadata {
  fn default() -> Self {
  Self {

indexify_extractor_sdk/base_embedding.py

@@ -15,7 +15,7 @@
 @dataclass
 class EmbeddingInputParams:
  overlap: int = 0
- chunk_size: int = 0
+ chunk_size: int = 100
  text_splitter: str = "recursive"
 
 
@@ -25,7 +25,7 @@ class BaseEmbeddingExtractor(Extractor):
  def __init__(self, max_context_length: int):
  self._model_context_length: int = max_context_length
 
- def extract(self, content: Content, params: EmbeddingInputParams) -> List[Content]:
+ def extract(self, content: Content, params: EmbeddingInputParams=EmbeddingInputParams()) -> List[Content]:
  if params.chunk_size == 0:
  params.chunk_size = self._model_context_length
  splitter: Callable[[str], List[str]] = self._create_splitter(params)

indexify_extractor_sdk/base_extractor.py

@@ -6,7 +6,6 @@
 
 from dataclasses import dataclass, field
 from dataclasses_json import dataclass_json
-from marshmallow_jsonschema import JSONSchema
 
 @dataclass_json
 @dataclass
@@ -143,7 +142,8 @@ def describe(self, input_params = None) -> ExtractorDescription:
  json_schema = {}
  json_schema["additionalProperties"] = False
  if self._param_cls and callable(getattr(self._param_cls, "schema", None)):
- json_schema = JSONSchema().dump(self._param_cls.schema())
+ # FIXME Construct the JSON Schema of the input params
+ json_schema = {}
  for content in out_c:
  for feature in content.features:
  if feature.feature_type == "embedding":

setup.py

@@ -33,7 +33,6 @@
  "transformers",
  "dataclasses-json",
  "marshmallow",
- "marshmallow-jsonschema",
 ]
 
 # What packages are optional?