V0.1.4 #4
V0.1.4 #4
Conversation
![mentatbot[bot]](https://avatars.githubusercontent.com/in/837875?s=60&v=4){kind=small}
| @@ -7,15 +7,15 @@ packages=["ragdaemon"] | |||
|
|
|||
| [project] | |||
| name = "ragdaemon" | |||
| version = "0.1.3" | |||
| version = "0.1.4" | |||
There was a problem hiding this comment.
Version bump in pyproject.toml matches the version update in ragdaemon/__init__.py. This ensures consistency across the project.
| @@ -1 +1 @@ | |||
| __version__ = "0.1.3" | |||
| __version__ = "0.1.4" | |||
There was a problem hiding this comment.
Correctly updated the version to 0.1.4. This change is necessary for version tracking and package management.
| from ragdaemon.annotators.chunker_line import ChunkerLine | ||
| from ragdaemon.annotators.chunker_llm import ChunkerLLM |
There was a problem hiding this comment.
Reordering imports does not affect functionality but adheres to the convention of organizing imports alphabetically. Good for readability and maintenance.
| @@ -3,15 +3,15 @@ | |||
| from pathlib import Path | |||
|
|
|||
| import networkx as nx | |||
| from spice import Spice | |||
There was a problem hiding this comment.
Adding from spice import Spice directly aligns with the removal of token_counter from ragdaemon.llm, centralizing token counting to the Spice client.
| @@ -2,11 +2,12 @@ | |||
| from pathlib import Path | |||
|
|
|||
| import networkx as nx | |||
| from spice import Spice | |||
There was a problem hiding this comment.
Importing Spice directly in this file, instead of using a custom token_counter function, simplifies the codebase by leveraging existing functionality from the Spice library.
| @@ -12,9 +12,9 @@ | |||
| from spice import Spice | |||
| from starlette.templating import Jinja2Templates | |||
|
|
|||
| from ragdaemon.llm import DEFAULT_COMPLETION_MODEL | |||
| from ragdaemon.database import DEFAULT_EMBEDDING_MODEL | |||
| from ragdaemon.daemon import Daemon | |||
There was a problem hiding this comment.
Removing unused imports and reordering them improves code cleanliness and readability.
| @@ -128,7 +128,9 @@ def get_context( | |||
| # TODO: Compare graph hashes, reconcile changes | |||
| context = context_builder | |||
| include_context_message = context.render() | |||
| include_tokens = token_counter(include_context_message) | |||
| include_tokens = self.spice_client.count_tokens( | |||
There was a problem hiding this comment.
Using self.spice_client.count_tokens directly instead of a custom token_counter function is a good refactor, leveraging the capabilities of the Spice client for token counting.
| @@ -1,10 +1,11 @@ | |||
| import os | |||
| from pathlib import Path | |||
|
|
|||
| from spice import Spice, SpiceError | |||
| from spice import Spice | |||
There was a problem hiding this comment.
Reordering imports and separating from spice.errors import SpiceError improves the clarity of where specific components are sourced from.
| @@ -1,12 +1 @@ | |||
| import tiktoken | |||
|
|
|||
|
|
|||
| DEFAULT_COMPLETION_MODEL = "gpt-4-0125-preview" | |||
There was a problem hiding this comment.
Removing the token_counter function and its dependencies simplifies the module, especially since token counting is now handled by the Spice client directly.
| @@ -6,7 +6,6 @@ | |||
|
|
|||
| from ragdaemon.errors import RagdaemonError | |||
|
|
|||
|
|
|||
| mentat_dir_path = Path.home() / ".mentat" | |||
There was a problem hiding this comment.
Removing unnecessary whitespace improves code cleanliness.
| @@ -5,7 +5,7 @@ Jinja2==3.1.3 | |||
| networkx==3.2.1 | |||
| pytest==8.0.2 | |||
| pytest-asyncio==0.23.5 | |||
| spiceai==0.1.9 | |||
| spiceai==0.1.11 | |||
There was a problem hiding this comment.
Updating the spiceai version in requirements.txt to 0.1.11 matches the version specified in pyproject.toml, ensuring consistency in dependency management.
|
MENTAT CODE REVIEW IN ACTIVE DEVELOPMENT. Only in use on mentat and internal repos. This pull request effectively updates the project version to |
No description provided.