Skip to content

10-minute Quickstart

This quickstart walks through install → model parsing → safe editing → dataset conversion.

1) Install

Follow Install.

2) Parse messy input into a typed model

>>> import omnipy as om

>>> readings = om.Model[list[int]]((101, '102', 103.0))
>>> readings

╭──────────────────╮
│ Model[list[int]] │
│                  │
│ [101, 102, 103]  │
╰──────────────────╯

3) Safe interactive manipulation

>>> try:
...     readings.append('invalid')
... except Exception as err:
...     print(type(err).__name__)
>>> readings
ValidationError

╭──────────────────╮
│ Model[list[int]] │
│                  │
│ [101, 102, 103]  │
╰──────────────────╯

4) Build a Dataset and batch-parse values

>>> batch = om.Dataset[om.Model[int]]({'sample_a': '1', 'sample_b': 2.0, 'sample_c': 3})
>>> batch

╭───┬────────────────┬────────────┬────────┬──────────────────╮
│ # │ Data file name │    Type    │ Length │ Size (in memory) │
│   │                │            │        │                  │
│ 0 │ sample_a       │ Model[int] │      - │        589 Bytes │
│ 1 │ sample_b       │ Model[int] │      - │        589 Bytes │
│ 2 │ sample_c       │ Model[int] │      - │        589 Bytes │
╰───┴────────────────┴────────────┴────────┴──────────────────╯

5) Convert nested records to pandas-ready tables

>>> records = om.JsonListOfDictsDataset({'rows': [{'id': 'a', 'value': '1'}, {'id': 'b', 'value': 2}]})
>>> records
>>> records_pd = records.to(om.PandasDataset)
>>> records_pd

╭───┬────────────────┬─────────────┬────────┬──────────────────╮
│ # │ Data file name │    Type     │ Length │ Size (in memory) │
│   │                │             │        │                  │
│ 0 │ rows           │ PandasModel │      2 │           9.5 kB │
╰───┴────────────────┴─────────────┴────────┴──────────────────╯

Next steps