Generate random samples from JSONL files by count or ratio.
npx skills add https://github.com/cas-bigdatalab/piflow --skill random-sampler
Please help me install this Agent Skill. Skill: random_sampler Source: https://github.com/cas-bigdatalab/piflow/tree/main/workspace/skills/random_sampler Command: npx skills add https://github.com/cas-bigdatalab/piflow --skill random-sampler
The random_sampler skill simplifies the task of sampling JSONL data, providing the ability to randomly select a subset of data with configurable count or ratio.
Use the random_sampler skill with --method count --count 50 --seed 123 to randomly select 50 records from a JSONL file with a random seed of 123.
High-intent search queries and answers about installing and using this skill.
You can perform random sampling on a JSONL file by specifying a desired count or ratio, which generates a representative data subset suitable for creating test datasets or training machine learning models.
Yes, data sampling supports count-based selection, allowing you to specify an exact number of records to randomly select from your JSONL file rather than relying solely on a ratio-based percentage.
You can ensure reproducible random sampling results by configuring a specific random seed, such as setting the seed parameter to a fixed integer, which guarantees the same records are selected across multiple runs.
No, this approach performs non-parametric random selection without the need for stratification, meaning it does not group data into distinct strata before sampling your JSONL records.
The best way to create a random test dataset is to apply non-parametric random sampling to your source JSONL file, extracting a representative data subset configured by either a fixed count or a specified ratio.
cas-bigdatalabPerform weighted random sampling on JSONL files with configurable weight fields.
hung-phanValidate ML datasets for schema integrity, drift, and quality with Pandera and Pydantic.
cas-bigdatalabConvert TSV files into JSONL format using Python.
cas-bigdatalabSample structured data by strata using pandas with specified sample sizes.
