support multiple keyterms for v2 listen client

[lynne-spark]

https://github.com/orgs/deepgram/discussions/1422#discussioncomment-14619854

Summary by CodeRabbit

• New Features

  • Support providing multiple keyterms when starting a Listen v2 streaming connection. Accepts a single keyterm or a list, enabling multiple keyterms in one request.

• Documentation

  • Updated parameter descriptions to reflect support for multiple keyterms and clarified character limit guidance.

[coderabbitai]

Walkthrough

The connect methods in V2Client/AsyncV2Client and RawV2Client/AsyncRawV2Client now accept keyterm as Optional[Union[str, List[str]]]. Query parameter construction branches to handle a single string or multiple strings, expanding lists into repeated keyterm parameters. Docstrings and type hints were updated accordingly.

Changes

Cohort / File(s)	Summary of Changes
Public V2 clients: sync and async src/deepgram/listen/v2/client.py	Updated connect signatures to accept keyterm: Optional[Union[str, List[str]]]. Added branching to build query params for single vs multiple keyterms. Updated docstrings to reflect union type and limits.
Raw V2 clients: sync and async src/deepgram/listen/v2/raw_client.py	Extended connect parameter keyterm to Optional[Union[str, List[str]]]. Implemented list vs string handling when constructing query parameters. Updated docstrings accordingly.

Sequence Diagram(s)

sequenceDiagram
  autonumber
  participant App as Caller
  participant Client as V2Client/AsyncV2Client
  participant Raw as RawV2Client/AsyncRawV2Client
  participant WS as WebSocket Endpoint

  App->>Client: connect(model, encoding, sample_rate, keyterm=... )
  Client->>Raw: connect(..., keyterm)
  alt keyterm is a list
    Raw->>Raw: For each item in keyterm list<br/>append query param keyterm=item
  else keyterm is a string
    Raw->>Raw: Append single query param keyterm=value
  else keyterm is None
    Raw->>Raw: Do not add keyterm param
  end
  Raw->>WS: Open websocket with query params
  WS-->>Raw: Connection established
  Raw-->>Client: Socket client instance
  Client-->>App: Socket client instance

Loading

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Pre-merge checks and finishing touches

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title Check	✅ Passed	The title succinctly and accurately captures the main change by stating support for multiple keyterms in the v2 listen client, matching the modifications to the connect method’s keyterm parameter without extraneous detail.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 0

🧹 Nitpick comments (2)

src/deepgram/listen/v2/raw_client.py (1)

92-96: Broaden to sequences (not just list) and skip empty entries (optional).

Current check only accepts list; tuples/other sequences won’t work. Also, empty strings become keyterm=. Consider accepting any non-string Sequence and skipping empties.

Apply this minimal change in both sync and async blocks:

-            if isinstance(keyterm, list):
+            from collections.abc import Sequence  # at top of file
+            if isinstance(keyterm, Sequence) and not isinstance(keyterm, (str, bytes)):
                 for kt in keyterm:
-                    query_params = query_params.add("keyterm", kt)
+                    if kt:  # skip empty strings
+                        query_params = query_params.add("keyterm", kt)
             else:
                 query_params = query_params.add("keyterm", keyterm)

Optional: enforce the documented 100‑char limit per keyterm with a simple length check before adding.

Also applies to: 197-201

src/deepgram/listen/v2/client.py (1)

104-108: Accept any non-string sequence and ignore empty values (optional).

Parity with raw client suggestion: support tuples/other sequences and avoid sending keyterm= for empty strings.

Apply in both sync and async connect paths:

-        if keyterm is not None:
-            if isinstance(keyterm, list):
+        if keyterm is not None:
+            from collections.abc import Sequence  # at top of file
+            if isinstance(keyterm, Sequence) and not isinstance(keyterm, (str, bytes)):
                 for kt in keyterm:
-                    query_params = query_params.add("keyterm", kt)
+                    if kt:
+                        query_params = query_params.add("keyterm", kt)
             else:
                 query_params = query_params.add("keyterm", keyterm)

Optionally validate each keyterm length (<=100) per docs.

Also applies to: 220-224

📜 Review details

Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 5a8c79e and e9b7794.

📒 Files selected for processing (2)

• src/deepgram/listen/v2/client.py (6 hunks)
• src/deepgram/listen/v2/raw_client.py (6 hunks)

🔇 Additional comments (2)

src/deepgram/listen/v2/raw_client.py (1)

34-34: Keyterm type/doc update looks good.

Union[str, List[str]] matches the intended multi-keyterm support; doc aligns.

Also applies to: 58-60

src/deepgram/listen/v2/client.py (1)

46-46: Keyterm type/doc update acknowledged.

Public surface now supports single or multiple keyterms; consistent with raw client.

Also applies to: 70-72

[lukeocodes]

@lynne-spark thanks for this, and sorry for the delay. I believe this will be resolved in #602 which I am hoping to release today. You could pin to that branch now to test it before it goes out Apologies, this isn't resolved in the upcoming release but it WILL be. We have another upstream change trickling out today before #602 is released

[lukeocodes]

Lets get this in and then we can figure out the upstream code generation after

[lynne-spark]

@lukeocodes just checking whether this is still in progress — I don’t see the fix in the latest SDK release.