codedoc-ai 0.7.0__tar.gz → 0.7.1__tar.gz

{codedoc_ai-0.7.0 → codedoc_ai-0.7.1}/CHANGELOG.md

@@ -1,5 +1,17 @@
 # Changelog
 
+## 0.7.1 - 2026-05-25
+
+### Fixed: provider-specific default models not applied when `--model` is omitted (GitHub Issue #2)
+
+- `codedoc/core/loader.py`: changed `DEFAULTS["model_name"]` from `"gpt-4o-mini"` to `""`.
+- Previously, the global default `"gpt-4o-mini"` was a truthy string that short-circuited the `or` fallbacks in the provider factory for every provider. Running `--provider gemini` without `--model` would silently send requests to Gemini using the OpenAI model name `gpt-4o-mini`, causing a 404 from the Gemini API. The same bug applied to `--provider anthropic` without `--model`, which would have called Anthropic with `gpt-4o-mini` and failed.
+- With an empty string default, the factory's per-provider fallbacks now activate correctly:
+  - Gemini with no model → `gemini-2.5-flash`
+  - Anthropic with no model → `claude-haiku-4-5-20251001`
+  - OpenAI / auto with no model → `gpt-4o-mini` (unchanged)
+- Behaviour when `--model` is explicitly passed is unchanged.
+
 ## 0.7.0 - 2026-05-24
 
 ### MD-only incremental now works (Issue 1)

{codedoc_ai-0.7.0/codedoc_ai.egg-info → codedoc_ai-0.7.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: codedoc-ai
-Version: 0.7.0
+Version: 0.7.1
 Summary: Local-first, LLM-agnostic codebase documentation for AI coding agents
 Author: Atharv Mannur
 License-Expression: MIT
@@ -40,7 +40,7 @@ Dynamic: license-file
 
 The tool scans source files, resolves project-local imports into a dependency graph, sends only files that need analysis to an LLM, and writes one combined, structured documentation artifact designed for both humans and AI. By default that artifact is JSON.
 
-Current release: `0.7.0`.
+Current release: `0.7.1`.
 
 ## What It Does
 

{codedoc_ai-0.7.0 → codedoc_ai-0.7.1}/README.md

@@ -4,7 +4,7 @@
 
 The tool scans source files, resolves project-local imports into a dependency graph, sends only files that need analysis to an LLM, and writes one combined, structured documentation artifact designed for both humans and AI. By default that artifact is JSON.
 
-Current release: `0.7.0`.
+Current release: `0.7.1`.
 
 ## What It Does
 

{codedoc_ai-0.7.0 → codedoc_ai-0.7.1}/codedoc/cli/cli.py

@@ -137,7 +137,7 @@ examples:
     parser.add_argument(
         "--version",
         action="version",
-        version="%(prog)s 0.7.0",
+        version="%(prog)s 0.7.1",
     )
 
     return parser

{codedoc_ai-0.7.0 → codedoc_ai-0.7.1}/codedoc/core/loader.py

@@ -21,7 +21,7 @@ logger = get_logger(__name__)
 DEFAULTS: dict[str, Any] = {
     "llm_mode": "api",
     "llm_provider": "auto",
-    "model_name": "gpt-4o-mini",
+    "model_name": "",
     "api_base_url": None,
     "api_key": None,
     "entry_file": None,

{codedoc_ai-0.7.0 → codedoc_ai-0.7.1/codedoc_ai.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: codedoc-ai
-Version: 0.7.0
+Version: 0.7.1
 Summary: Local-first, LLM-agnostic codebase documentation for AI coding agents
 Author: Atharv Mannur
 License-Expression: MIT
@@ -40,7 +40,7 @@ Dynamic: license-file
 
 The tool scans source files, resolves project-local imports into a dependency graph, sends only files that need analysis to an LLM, and writes one combined, structured documentation artifact designed for both humans and AI. By default that artifact is JSON.
 
-Current release: `0.7.0`.
+Current release: `0.7.1`.
 
 ## What It Does
 

{codedoc_ai-0.7.0 → codedoc_ai-0.7.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "codedoc-ai"
-version = "0.7.0"
+version = "0.7.1"
 description = "Local-first, LLM-agnostic codebase documentation for AI coding agents"
 readme = "README.md"
 license = "MIT"