fix: correct TextQuery and VectorQuery API serialization

[jhamon]

Summary

Fixes unresolved review comments from PR #586 that were merged without being addressed.

Related PR: #586

Problems Fixed

1. Wrong API field name in TextQuery (High Severity)

The as_dict() method was serializing the query string with key "query", but the Pinecone API expects "text_query" as defined in the OpenAPI spec.

2. Missing type discriminator field (High Severity)

Both TextQuery and VectorQuery were missing the required type discriminator field used by the ScoreByQuery oneOf schema:

• TextQuery → "type": "text"
• VectorQuery → "type": "vector"

3. Inconsistent method naming (Low Severity)

Changed to_dict() to as_dict() for consistency with similar search-related classes (SearchQuery, SearchQueryVector, SearchRerank).

Breaking Changes

Since these classes were just added in PR #586 and haven't been released yet, these changes should not affect any users:

• TextQuery.to_dict() → TextQuery.as_dict()
• VectorQuery.to_dict() → VectorQuery.as_dict()

Usage Example

from pinecone import TextQuery, VectorQuery

# TextQuery for full-text search
text_query = TextQuery(field="title", query="pink panther")
text_query.as_dict()
# Returns: {"type": "text", "field": "title", "text_query": "pink panther"}

# With optional parameters
text_query = TextQuery(field="title", query="pink panther", boost=2.0, slop=1)
text_query.as_dict()
# Returns: {"type": "text", "field": "title", "text_query": "pink panther", "boost": 2.0, "slop": 1}

# VectorQuery for similarity search
vector_query = VectorQuery(field="embedding", values=[0.1, 0.2, 0.3])
vector_query.as_dict()
# Returns: {"type": "vector", "field": "embedding", "values": [0.1, 0.2, 0.3]}

# Use with search_documents()
results = index.search_documents(
    namespace="movies",
    score_by=TextQuery(field="title", query="pink panther"),
    top_k=10,
)

Test Plan

• Updated unit tests to verify correct API format with type discriminator
• Updated tests to use as_dict() method name
• All 16 query class tests pass
• mypy type checking passes

---

Note

Medium Risk
Medium risk because it changes the serialized request shape and renames a public method (to_dict() → as_dict()), which can break any existing callers even though the change is localized.

Overview
Fixes TextQuery and VectorQuery API serialization used by score_by by adding the required type discriminator ("text"/"vector") and aligning field names (notably query → text_query for text scoring).

Renames the serialization helper from to_dict() to as_dict() on both classes and updates unit tests to assert the new payload format and method name.

^{Written by Cursor Bugbot for commit 50ee953. This will update automatically on new commits. Configure here.}