A CLI tool for discovering similar notes within your note collection.
Project description
SimiNotes - CLI Tool for Similar Note Retrieval
SimiNotes is a command-line interface (CLI) tool written in Python that enables users to discover similar notes within their notes collection. The tool utilizes sentence embeddings with sbert to compare a given query against a corpus of user notes.
Features
- Embedding: Utilizes sbert to generate vector embeddings for notes.
- Similarity Search: Finds notes similar to a given query based on embeddings.
- Configurable: Allows users to configure directories, file extensions, and exclusion criteria.
Installation
Before using this CLI tool, ensure that you have Python and pip installed. Additionally, install the PyTorch library by following the steps below:
-
Download and install Python and pip from the official site.
-
Install the PyTorch library separately based on your requirements. For the CPU version, use the following command:
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
Now, install the SimiNotes CLI tool:
pip install siminotes
Configuration
Before using the CLI, configure some essential values, such as the notes directory and exclusions. SimiNotes uses a
configuration file (config.txt
) to set preferences. Configure the tool by creating the file in the appropriate configuration
directory.
Configuration Directory
- Linux:
~/.config/siminotesconfig/
- macOS:
~/.siminotesconfig/
- Windows:
For Windows, use AppData\Roaming for per-user configuration:
AppData\Roaming\Siminotes\
Alternatively, place config.txt
in the home directory:
~/.siminotesconfig/
Configuration File (config.txt
)
Create a config.txt
file in the configuration directory. Below is an example configuration:
notes_dir = /path/to/your/notes
exclude_dir = directory1,directory2
exclude_file = file1,file2
note_extension = .md
Configuration Parameters:
-
notes_dir: Path to the directory containing your notes.
-
exclude_dir: Comma-separated directories to exclude from the search. Paths should be relative to notes_dir.
-
exclude_file: Comma-separated files to exclude from the search. Paths should be relative to notes_dir.
-
note_extension: The extension of your note files (e.g., .md).
Usage
Now let's use our cli,
Command-Line Arguments
- Query via Text:
siminotes text "Your Query Text"
- Query via File:
siminotes file filename
Both will result in,
Top files which are similar to given query:
Value range from -1 to 1, where going toward 1 means note is close to query
/... with score 0.43386968970298767
/... with score 0.42138463258743286
...
Troubleshooting
If you encounter any errors or problems with this tool, please open an issue in the repository.
License
This project is licensed under the MIT License.
Contributing
Feel free to contribute to SimiNotes by creating issues or submitting pull requests.
Acknowledgments
Sentence-BERT for sentence embeddings.
Future:
-
I feel like simple dot product hits are enough to find similar notes but in future if there is need to improve results then consider this roop Retrieve and Rerank Vid Tut
-
If it is taking more memory, then we can quantise the vectors into int8 Quantisation Guide Github Repo to check
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for siminotes-0.1.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 00744fc1f4e8aee812108661a11bda2631272c9cdfa7a9a645dcb30774a72dc7 |
|
MD5 | 95aa42cddba121fe38e39ca7a67762fd |
|
BLAKE2b-256 | 95ac2d59820942dc74422248cf9887393933df6f5dc32edede0326236e8388b7 |