Usage
SeaGOAT is a command-line tool designed to assist in querying your codebase. By using technologies such as ChromaDB and ripgrep, it goes beyond direct match searches and uses semantic meaning to quickly find details related to your query.
Only works with Git
SeaGOAT takes your Git history into account in order to provide the most useful and relevant results.
Command Usage
Note
The seagoat CLI queries the SeaGOAT server. If the server is not running,
you would be prompted to start the server using
seagoat-server start {repo_path} command.
Arguments
query: This is a required argument. It is the query to be made to the SeaGOAT server.repo_path: This argument is optional, and defaults to the current working directory. It represents the path to the code repository.
Examples
Query current folder
Query specific folder
Using Regular Expressions
One of SeaGOAT's most powerful features is the ability to combine regular expressions with AI-driven vector queries. This synergistic approach narrows down your codebase search using pattern-based regular expressions while leveraging AI to understand the semantic meaning behind your query.
Options
--no-color: Disable syntax highlighting
This is automatically enabled when used as part of a bash pipeline.
-l, --max-results: Limit number of result lines
This limits the number of result lines displayed. Useful if you only care about the best results.
SeaGOAT is oriented around code blocks, not individual lines
In SeaGOAT, code is displayed in full, continous blocks rather than individual lines. It'll always show at least one full block, even if your limit is 0.
If you set a limit, SeaGOAT ensures that complete blocks are shown as long as they fit in your limit. For example, with a 5-line limit, it can show a 3-line and a 2-line block, but not two 3-line blocks.
Also, this limit only counts the actual code, not any extra context lines you might request.
--version: Print version number
This prints the version of your current SeaGOAT installation.
-B, --context-above: Lines of context before each result
This option allows you to include a specified number of lines of context before each matching result.
Tricky context lines
Context lines are lines that are added because they are adjacent to a result line.
That being said, because lines are grouped into chunks of 3, results based on vector embeddings might already contain lines that might not be strictly related to the query.
This might make it appear like there are more context lines than you requested. Consider this when deciding how many context lines to include.
-A, --context-below: Lines of context after each result
This option allows you to include a specified number of lines of context after each matching result.
Tricky context lines
Context lines are lines that are added because they are adjacent to a result line.
That being said, because lines are grouped into chunks of 3, results based on vector embeddings might already contain lines that might not be strictly related to the query.
This might make it appear like there are more context lines than you requested. Consider this when deciding how many context lines to include.
-C, --context: Lines of context both before and after each result
This option sets both --context-above and --context-below to the same
specified value. This is useful if you want an equal amount of context around
each matching result.
Tricky context lines
Context lines are lines that are added because they are adjacent to a result line.
That being said, because lines are grouped into chunks of 3, results based on vector embeddings might already contain lines that might not be strictly related to the query.
This might make it appear like there are more context lines than you requested. Consider this when deciding how many context lines to include.