tsugite-codex-cli 0.18.0__tar.gz → 0.19.0__tar.gz

{tsugite_codex_cli-0.18.0 → tsugite_codex_cli-0.19.0}/.gitignore

@@ -15,8 +15,8 @@ dist/
 downloads/
 eggs/
 .eggs/
-lib/
-lib64/
+/lib/
+/lib64/
 parts/
 sdist/
 var/
@@ -176,9 +176,21 @@ docs-old/
 docs/design/
 examples/*
 !examples/tsugite-example-plugin/
-agents/
+/agents/
 .claude/
 .tsugite/
 benchmarks/
 docker-compose.test.yml
-#### TODO ^^^
+#### TODO ^^^
+
+plugins/tsugite-daemon/frontend/node_modules/
+# Vite dep-optimizer cache. Normally under node_modules/.vite, but lands at the
+# cwd's `.vite/` when vite/vitest is run from the repo root, so ignore it anywhere.
+.vite/
+# Built web UI - generated by `mise run web-build`; the wheel bundles it via
+# hatch `artifacts`, git never carries it.
+plugins/tsugite-daemon/tsugite_daemon/web/
+# Re-include frontend source that generic lib/ rules (incl. contributors' global
+# gitignores with Python-style `lib/`) would otherwise silently exclude.
+!plugins/tsugite-daemon/frontend/src/lib/
+!plugins/tsugite-daemon/frontend/src/lib/**

{tsugite_codex_cli-0.18.0 → tsugite_codex_cli-0.19.0}/PKG-INFO

@@ -1,7 +1,7 @@
 Metadata-Version: 2.4
 Name: tsugite-codex-cli
-Version: 0.18.0
+Version: 0.19.0
 Summary: Tsugite plugin: ChatGPT subscription auth via Codex CLI token store
 Author: Justyn Shull
 Requires-Python: >=3.11
-Requires-Dist: tsugite-cli==0.18.0
+Requires-Dist: tsugite-cli==0.19.0

{tsugite_codex_cli-0.18.0 → tsugite_codex_cli-0.19.0}/pyproject.toml

@@ -1,11 +1,11 @@
 [project]
 name = "tsugite-codex-cli"
-version = "0.18.0"
+version = "0.19.0"
 description = "Tsugite plugin: ChatGPT subscription auth via Codex CLI token store"
 authors = [{ name = "Justyn Shull" }]
 requires-python = ">=3.11"
 dependencies = [
-  "tsugite-cli==0.18.0",
+  "tsugite-cli==0.19.0",
 ]
 
 [project.entry-points."tsugite.providers"]

{tsugite_codex_cli-0.18.0 → tsugite_codex_cli-0.19.0}/tsugite_codex_cli/provider.py

@@ -88,7 +88,7 @@ def _normalise_kwargs(kwargs: dict) -> dict:
         if key == "max_tokens":
             body["max_output_tokens"] = value
         elif key == "reasoning_effort":
-            body["reasoning"] = {"effort": value}
+            body["reasoning"] = {"effort": value, "summary": "auto"}
         elif key == "response_format":
             body["text"] = {"format": value}
         elif key in {"temperature", "top_p", "stop", "presence_penalty", "frequency_penalty"}:
@@ -107,10 +107,16 @@ def _build_usage(usage_in: dict | None) -> Usage:
     total_tokens = int(total) if total is not None else input_tokens + output_tokens
     details = usage_in.get("output_tokens_details") or {}
     reasoning_tokens = details.get("reasoning_tokens")
+    # The Responses API reports the cached prompt prefix (a subset of input_tokens)
+    # under input_tokens_details.cached_tokens - the input-side mirror of
+    # output_tokens_details.reasoning_tokens above.
+    input_details = usage_in.get("input_tokens_details") or {}
+    cached_tokens = input_details.get("cached_tokens")
     return Usage(
         prompt_tokens=input_tokens,
         completion_tokens=output_tokens,
         total_tokens=total_tokens,
+        cached_tokens=int(cached_tokens) if cached_tokens is not None else None,
         reasoning_tokens=int(reasoning_tokens) if reasoning_tokens is not None else None,
     )
 
@@ -120,6 +126,7 @@ class CodexCliProvider:
 
     # Stateless across calls (auth refresh is internal), so the registry can cache one instance.
     cacheable = True
+    models_are_definitive = True  # CLI exposes a fixed model set; unlisted ids are typos
 
     def __init__(self, name: str = "codex_cli"):
         self.name = name
@@ -178,6 +185,11 @@ class CodexCliProvider:
         if stream:
             body["stream"] = True
         body.update(_normalise_kwargs(kwargs))
+        # Reasoning summaries are opt-in on the Responses API: without
+        # `reasoning.summary` no reasoning_summary_text deltas are ever sent,
+        # so the UI can never show thinking. Ask for them even when no effort
+        # override is set (the codex CLI itself always sends both).
+        body.setdefault("reasoning", {}).setdefault("summary", "auto")
         return body
 
     @staticmethod
@@ -279,6 +291,10 @@ class CodexCliProvider:
                     delta = event.get("delta") or ""
                     if delta:
                         yield StreamChunk(reasoning_content=delta)
+                elif etype == "response.reasoning_summary_part.done":
+                    # Summary parts are separate paragraphs; without a break the
+                    # accumulated markdown runs them together (**a****b**).
+                    yield StreamChunk(reasoning_content="\n\n")
                 elif etype == "response.completed":
                     response_obj = event.get("response") or {}
                     usage = _build_usage(response_obj.get("usage"))