datasette-rure

Datasette plugin that adds a custom SQL function for executing matches using the Rust regular expression engine

Install this plugin in the same environment as Datasette to enable the regexp() SQL function.

$ pip install datasette-rure

The plugin is built on top of the rure-python library by David Blewett.

regexp() to test regular expressions

You can test if a value matches a regular expression like this:

select regexp('hi.*there', 'hi there')
-- returns 1
select regexp('not.*there', 'hi there')
-- returns 0

You can also use SQLite's custom syntax to run matches:

select 'hi there' REGEXP 'hi.*there'
-- returns 1

This means you can select rows based on regular expression matches - for example, to select every article where the title begins with an E or an F:

select * from articles where title REGEXP '^[EF]'

Try this out: REGEXP interactive demo

regexp_match() to extract groups

You can extract captured subsets of a pattern using regexp_match().

select regexp_match('.*( and .*)', title) as n from articles where n is not null
-- Returns the ' and X' component of any matching titles, e.g.
--     and Recognition
--     and Transitions Their Place
-- etc

This will return the first parenthesis match when called with two arguments. You can call it with three arguments to indicate which match you would like to extract:

select regexp_match('.*(and)(.*)', title, 2) as n from articles where n is not null

The function will return null for invalid inputs e.g. a pattern without capture groups.

Try this out: regexp_match() interactive demo

regexp_matches() to extract multiple matches at once

The regexp_matches() function can be used to extract multiple patterns from a single string. The result is returned as a JSON array, which can then be further processed using SQLite's JSON functions.

The first argument is a regular expression with named capture groups. The second argument is the string to be matched.

select regexp_matches(
    'hello (?P<name>\w+) the (?P<species>\w+)',
    'hello bob the dog, hello maggie the cat, hello tarquin the otter'
)

This will return a list of JSON objects, each one representing the named captures from the original regular expression:

[
    {""name"": ""bob"", ""species"": ""dog""},
    {""name"": ""maggie"", ""species"": ""cat""},
    {""name"": ""tarquin"", ""species"": ""otter""}
]

Try this out: regexp_matches() interactive demo

datasette-render-timestamps

Datasette plugin for rendering timestamps.

Installation

Install this plugin in the same environment as Datasette to enable this new functionality:

pip install datasette-render-timestamps

The plugin will then look out for integer numbers that are likely to be timestamps - anything that would be a number of seconds from 5 years ago to 5 years in the future.

These will then be rendered in a more readable format.

Configuration

You can disable automatic column detection in favour of explicitly listing the columns that you would like to render using plugin configuration in a metadata.json file.

Add a ""datasette-render-timestamps"" configuration block and use a ""columns"" key to list the columns you would like to treat as timestamp values:

{
    ""plugins"": {
        ""datasette-render-timestamps"": {
            ""columns"": [""created"", ""updated""]
        }
    }
}

This will cause any created or updated columns in any table to be treated as timestamps and rendered.

Save this to metadata.json and run datasette with the --metadata flag to load this configuration:

datasette serve mydata.db --metadata metadata.json

To disable automatic timestamp detection entirely, you can use ""columnns"": [].

This configuration block can be used at the top level, or it can be applied just to specific databases or tables. Here's how to apply it to just the entries table in the news.db database:

{
    ""databases"": {
        ""news"": {
            ""tables"": {
                ""entries"": {
                    ""plugins"": {
                        ""datasette-render-timestamps"": {
                            ""columns"": [""created"", ""updated""]
                        }
                    }
                }
            }
        }
    }
}

And here's how to apply it to every created column in every table in the news.db database:

{
    ""databases"": {
        ""news"": {
            ""plugins"": {
                ""datasette-render-timestamps"": {
                    ""columns"": [""created"", ""updated""]
                }
            }
        }
    }
}

Customizing the date format

The default format is %B %d, %Y - %H:%M:%S UTC which renders for example: October 10, 2019 - 07:18:29 UTC. If you want another format, the date format can be customized using plugin configuration. Any format string supported by strftime may be used. For example:

{
    ""plugins"": {
        ""datasette-render-timestamps"": {
            ""format"": ""%Y-%m-%d-%H:%M:%S""
        }
    }
}