feat(agents-api): Performance improvements (#749)

- **feat(agents-api,typespec): Limit max_k for search to 50; max items
for embed to 100**
- **refactor(typespec): Remove older 0.4.0 version of typespec**
- **feat: Add metadata filter argument to doc search**
- **feat(agents-api): Performance improvements - use KNN unless dataset
too big**
- **feat(agents-api): Performance improvements - merge metadata_filter**

<!-- ELLIPSIS_HIDDEN -->

----

> [!IMPORTANT]
> This PR improves agents API performance by limiting search and embed
parameters, adding metadata filtering, updating typespec versions, and
optimizing search functions.
> 
>   - **Behavior**:
> - Limit `max_k` for search to 50 and max items for embed to 100 in
`Docs.py`.
> - Add `metadata_filter` argument to document search functions in
`search_docs_by_embedding.py`, `search_docs_by_text.py`, and
`search_docs_hybrid.py`.
> - Use KNN for search unless dataset exceeds `ann_threshold` in
`search_docs_by_embedding.py`.
>   - **Refactor**:
>     - Remove older `0.4.0` version of typespec.
>     - Update `openapi.yaml` to `1.0.0` version.
>   - **Models**:
> - Split `EmbedQueryRequest` into `SingleEmbedQueryRequest` and
`MultipleEmbedQueryRequest` in `Docs.py` and `models.tsp`.
>   - **Misc**:
>     - Adjust `workers` and add `timeout` in `gunicorn_conf.py`.
>     - Update `docker-compose.yml` to sync `gunicorn_conf.py`.
> 
> <sup>This description was created by </sup>[<img alt="Ellipsis"
src="https://img.shields.io/badge/Ellipsis-blue?color=175173">](https://www.ellipsis.dev?ref=julep-ai%2Fjulep&utm_source=github&utm_medium=referral)<sup>
for c0bd25c. It will automatically
update as commits are pushed.</sup>

<!-- ELLIPSIS_HIDDEN -->

---------

Signed-off-by: Diwank Singh Tomer <[email protected]>
Co-authored-by: Dmitry Paramonov <[email protected]>

agents-api/agents_api/autogen/Agents.py

@@ -1,5 +1,5 @@
 # generated by datamodel-codegen:
-#   filename:  openapi-0.4.0.yaml
+#   filename:  openapi-1.0.0.yaml
 
 from __future__ import annotations
 

agents-api/agents_api/autogen/Chat.py

@@ -1,5 +1,5 @@
 # generated by datamodel-codegen:
-#   filename:  openapi-0.4.0.yaml
+#   filename:  openapi-1.0.0.yaml
 
 from __future__ import annotations
 

agents-api/agents_api/autogen/Common.py

@@ -1,5 +1,5 @@
 # generated by datamodel-codegen:
-#   filename:  openapi-0.4.0.yaml
+#   filename:  openapi-1.0.0.yaml
 
 from __future__ import annotations
 

agents-api/agents_api/autogen/Docs.py

@@ -1,23 +1,24 @@
 # generated by datamodel-codegen:
-#   filename:  openapi-0.4.0.yaml
+#   filename:  openapi-1.0.0.yaml
 
 from __future__ import annotations
 
 from typing import Annotated, Any, Literal
 from uuid import UUID
 
-from pydantic import AwareDatetime, BaseModel, ConfigDict, Field
+from pydantic import AwareDatetime, BaseModel, ConfigDict, Field, StrictBool
 
 
 class BaseDocSearchRequest(BaseModel):
     model_config = ConfigDict(
         populate_by_name=True,
     )
-    limit: Annotated[int, Field(10, ge=1, le=100)]
+    limit: Annotated[int, Field(10, ge=1, le=50)]
     lang: Literal["en-US"] = "en-US"
     """
     The language to be used for text-only search. Support for other languages coming soon.
     """
+    metadata_filter: dict[str, float | str | StrictBool | None] = {}
 
 
 class CreateDocRequest(BaseModel):
@@ -109,20 +110,6 @@ class DocSearchResponse(BaseModel):
     """
 
 
-class EmbedQueryRequest(BaseModel):
-    model_config = ConfigDict(
-        populate_by_name=True,
-    )
-    text: str | list[str]
-    """
-    Text or texts to embed
-    """
-    embed_instruction: str = ""
-    """
-    Instruction for the embedding model.
-    """
-
-
 class EmbedQueryResponse(BaseModel):
     model_config = ConfigDict(
         populate_by_name=True,
@@ -155,6 +142,34 @@ class HybridDocSearchRequest(BaseDocSearchRequest):
     """
 
 
+class MultipleEmbedQueryRequest(BaseModel):
+    model_config = ConfigDict(
+        populate_by_name=True,
+    )
+    text: Annotated[list[str], Field(max_length=100, min_length=1)]
+    """
+    Texts to embed
+    """
+    embed_instruction: str = ""
+    """
+    Instruction for the embedding model.
+    """
+
+
+class SingleEmbedQueryRequest(BaseModel):
+    model_config = ConfigDict(
+        populate_by_name=True,
+    )
+    text: str
+    """
+    Text to embed
+    """
+    embed_instruction: str = ""
+    """
+    Instruction for the embedding model.
+    """
+
+
 class Snippet(BaseModel):
     model_config = ConfigDict(
         populate_by_name=True,

agents-api/agents_api/autogen/Entries.py

@@ -1,5 +1,5 @@
 # generated by datamodel-codegen:
-#   filename:  openapi-0.4.0.yaml
+#   filename:  openapi-1.0.0.yaml
 
 from __future__ import annotations
 

agents-api/agents_api/autogen/Executions.py

@@ -1,5 +1,5 @@
 # generated by datamodel-codegen:
-#   filename:  openapi-0.4.0.yaml
+#   filename:  openapi-1.0.0.yaml
 
 from __future__ import annotations
 

agents-api/agents_api/autogen/Jobs.py

@@ -1,5 +1,5 @@
 # generated by datamodel-codegen:
-#   filename:  openapi-0.4.0.yaml
+#   filename:  openapi-1.0.0.yaml
 
 from __future__ import annotations
 

agents-api/agents_api/autogen/Sessions.py

@@ -1,5 +1,5 @@
 # generated by datamodel-codegen:
-#   filename:  openapi-0.4.0.yaml
+#   filename:  openapi-1.0.0.yaml
 
 from __future__ import annotations
 

agents-api/agents_api/autogen/Tasks.py

@@ -1,5 +1,5 @@
 # generated by datamodel-codegen:
-#   filename:  openapi-0.4.0.yaml
+#   filename:  openapi-1.0.0.yaml
 
 from __future__ import annotations
 

agents-api/agents_api/autogen/Tools.py

@@ -1,5 +1,5 @@
 # generated by datamodel-codegen:
-#   filename:  openapi-0.4.0.yaml
+#   filename:  openapi-1.0.0.yaml
 
 from __future__ import annotations
 

agents-api/agents_api/autogen/Users.py

@@ -1,5 +1,5 @@
 # generated by datamodel-codegen:
-#   filename:  openapi-0.4.0.yaml
+#   filename:  openapi-1.0.0.yaml
 
 from __future__ import annotations
 

agents-api/agents_api/autogen/__init__.py

@@ -1,2 +1,2 @@
 # generated by datamodel-codegen:
-#   filename:  openapi-0.4.0.yaml
+#   filename:  openapi-1.0.0.yaml