Make some structural changes to how information is delivered. It has …

…a consistent pattern now and logically separates schema from signatures and examples, but still indicates which are for class descriptions and which are for instance descriptions

adalflow/adalflow/core/base_data_class.py

@@ -114,19 +114,24 @@ class DataClass:
 
     Describing:
 
-    We defined :class:`DataClassFormatType<core.types.DataClassFormatType>` to categorize DataClass description formats
-    as inputs or outputs in an LLM prompt.
+    We defined :class:`DataClassFormatType<core.types.DataClassFormatType>` to categorize DataClass formats
+    as LLM inputs or outputs. This can be broken down into:
+    (1) schema (class description)
+    (2) signatures (class description)
+    (3) examples (instance description)
 
+    (1) `DataClassFormatType.SCHEMA`
+    - Standard JSON-based desription, via: :meth:`to_schema` as string and :meth:`to_schema` as dict.
 
-    (1) For describing the class (data structure):
+    (2) `DataClassFormatType.SIGNATURE_JSON` / `DataClassFormatType.SIGNATURE_YAML`
+    - More token-efficient than SCHEMA. Since SCHEMA is always represented as a JSON string, describing the data structure in JSON may be misleading when you want LLMS to output YAML.
 
-    `Signature` is more token-effcient than schema. Since schema is always represented as a JSON string, describing the data structure in JSON may be misleading when you want LLMS to output YAML.
-
-    - DataClassFormatType.SCHEMA: a more standard way to describe the data structure in Json string, :meth:`to_schema` as string and :meth:`to_schema` as dict.
     - DataClassFormatType.SIGNATURE_JSON: imitating a json object with field name as key and description as value, :meth:`to_json_signature` as string.
     - DataClassFormatType.SIGNATURE_YAML: imitating a yaml object with field name as key and description as value, :meth:`to_yaml_signature` as string.
 
-    (2) For describing the class instance: this is helpful to do few-shot examples in LLM prompts.
+    (3) `DataClassFormatType.EXAMPLE_JSON` / `DataClassFormatType.EXAMPLE_YAML`
+    - Helpful to do few-shot examples in LLM prompts.
+
     - DataClassFormatType.EXAMPLE_JSON: the json representation of the instance, :meth:`to_json` as string.
     - DataClassFormatType.EXAMPLE_YAML: the yaml representation of the instance, :meth:`to_yaml` as string.