| chore: drop GitHub Actions and internal experiment notes Co-authored-by: Cursor <cursoragent@cursor.com> | 2 个月前 |
| chore: drop leftover experiment scripts and deprecated aliases Co-authored-by: Cursor <cursoragent@cursor.com> | 2 个月前 |
| refactor(taoctx)!: rename package and CLI to taoctx, ship runnable examples The project positioning is context orchestration rather than memory, so the package moves from taomem to taoctx and the CLI from taomem-serve to taoctx-serve. The TAOMEM_* env prefix and benchmark field names keep their historical spelling so existing experiment scripts and result schemas stay valid. - Move the OpenAI-compatible proxy from examples/api_server.py into src/taoctx/api/server.py; taoctx-serve previously failed after a plain install because examples/ is not part of the distribution. - Add beaver/prompt.py::compress_prompt plus examples/compress_block.py for the block-compression-only path, and keep it importable without torch by deferring the runtime import. - Add examples/sdk_gateway.py and examples/openai_compatible_client.py, and give the LangGraph / LangChain examples a --prepare-only mode so they run without a main LLM. - Restore TAOMEM_* / RETRIEVAL_* env between tests to stop cross-test leakage. | 2 个月前 |
| docs: add setup guides and scrub intermediate wording Add getting_started/tuning and model download links, prefer repo-local models/ paths, and remove WIP/task-style phrasing from docs and public module notes. Co-authored-by: Cursor <cursoragent@cursor.com> | 2 个月前 |
| chore: drop GitHub Actions and internal experiment notes Co-authored-by: Cursor <cursoragent@cursor.com> | 2 个月前 |
| chore: drop leftover experiment scripts and deprecated aliases Co-authored-by: Cursor <cursoragent@cursor.com> | 2 个月前 |
| chore: drop leftover experiment scripts and deprecated aliases Co-authored-by: Cursor <cursoragent@cursor.com> | 2 个月前 |
| chore: drop OpenClaw deploy and beaver run reports Co-authored-by: Cursor <cursoragent@cursor.com> | 2 个月前 |
| chore: drop OpenClaw deploy and beaver run reports Co-authored-by: Cursor <cursoragent@cursor.com> | 2 个月前 |
| refactor(taoctx)!: rename package and CLI to taoctx, ship runnable examples The project positioning is context orchestration rather than memory, so the package moves from taomem to taoctx and the CLI from taomem-serve to taoctx-serve. The TAOMEM_* env prefix and benchmark field names keep their historical spelling so existing experiment scripts and result schemas stay valid. - Move the OpenAI-compatible proxy from examples/api_server.py into src/taoctx/api/server.py; taoctx-serve previously failed after a plain install because examples/ is not part of the distribution. - Add beaver/prompt.py::compress_prompt plus examples/compress_block.py for the block-compression-only path, and keep it importable without torch by deferring the runtime import. - Add examples/sdk_gateway.py and examples/openai_compatible_client.py, and give the LangGraph / LangChain examples a --prepare-only mode so they run without a main LLM. - Restore TAOMEM_* / RETRIEVAL_* env between tests to stop cross-test leakage. | 2 个月前 |
| docs: add setup guides and scrub intermediate wording Add getting_started/tuning and model download links, prefer repo-local models/ paths, and remove WIP/task-style phrasing from docs and public module notes. Co-authored-by: Cursor <cursoragent@cursor.com> | 2 个月前 |
| chore: drop leftover experiment scripts and deprecated aliases Co-authored-by: Cursor <cursoragent@cursor.com> | 2 个月前 |
| docs: add setup guides and scrub intermediate wording Add getting_started/tuning and model download links, prefer repo-local models/ paths, and remove WIP/task-style phrasing from docs and public module notes. Co-authored-by: Cursor <cursoragent@cursor.com> | 2 个月前 |
| refactor(taoctx)!: rename package and CLI to taoctx, ship runnable examples The project positioning is context orchestration rather than memory, so the package moves from taomem to taoctx and the CLI from taomem-serve to taoctx-serve. The TAOMEM_* env prefix and benchmark field names keep their historical spelling so existing experiment scripts and result schemas stay valid. - Move the OpenAI-compatible proxy from examples/api_server.py into src/taoctx/api/server.py; taoctx-serve previously failed after a plain install because examples/ is not part of the distribution. - Add beaver/prompt.py::compress_prompt plus examples/compress_block.py for the block-compression-only path, and keep it importable without torch by deferring the runtime import. - Add examples/sdk_gateway.py and examples/openai_compatible_client.py, and give the LangGraph / LangChain examples a --prepare-only mode so they run without a main LLM. - Restore TAOMEM_* / RETRIEVAL_* env between tests to stop cross-test leakage. | 2 个月前 |
| refactor(taoctx)!: rename package and CLI to taoctx, ship runnable examples The project positioning is context orchestration rather than memory, so the package moves from taomem to taoctx and the CLI from taomem-serve to taoctx-serve. The TAOMEM_* env prefix and benchmark field names keep their historical spelling so existing experiment scripts and result schemas stay valid. - Move the OpenAI-compatible proxy from examples/api_server.py into src/taoctx/api/server.py; taoctx-serve previously failed after a plain install because examples/ is not part of the distribution. - Add beaver/prompt.py::compress_prompt plus examples/compress_block.py for the block-compression-only path, and keep it importable without torch by deferring the runtime import. - Add examples/sdk_gateway.py and examples/openai_compatible_client.py, and give the LangGraph / LangChain examples a --prepare-only mode so they run without a main LLM. - Restore TAOMEM_* / RETRIEVAL_* env between tests to stop cross-test leakage. | 2 个月前 |
| refactor(taoctx)!: rename package and CLI to taoctx, ship runnable examples The project positioning is context orchestration rather than memory, so the package moves from taomem to taoctx and the CLI from taomem-serve to taoctx-serve. The TAOMEM_* env prefix and benchmark field names keep their historical spelling so existing experiment scripts and result schemas stay valid. - Move the OpenAI-compatible proxy from examples/api_server.py into src/taoctx/api/server.py; taoctx-serve previously failed after a plain install because examples/ is not part of the distribution. - Add beaver/prompt.py::compress_prompt plus examples/compress_block.py for the block-compression-only path, and keep it importable without torch by deferring the runtime import. - Add examples/sdk_gateway.py and examples/openai_compatible_client.py, and give the LangGraph / LangChain examples a --prepare-only mode so they run without a main LLM. - Restore TAOMEM_* / RETRIEVAL_* env between tests to stop cross-test leakage. | 2 个月前 |
| docs: add setup guides and scrub intermediate wording Add getting_started/tuning and model download links, prefer repo-local models/ paths, and remove WIP/task-style phrasing from docs and public module notes. Co-authored-by: Cursor <cursoragent@cursor.com> | 2 个月前 |
