tanat.store.source package

Subpackages

• tanat.store.source.type package
  • Submodules
  • tanat.store.source.type.dataframe module
    • DataFrameSource
      • DataFrameSource.__init__()
      • DataFrameSource.read()
      • DataFrameSource.schema()
  • tanat.store.source.type.file module
    • CsvSource
      • CsvSource.__init__()
      • CsvSource.read()
      • CsvSource.schema()
    • ParquetSource
      • ParquetSource.__init__()
      • ParquetSource.read()
      • ParquetSource.schema()
  • tanat.store.source.type.sql module
    • SqlSource
      • SqlSource.__init__()
      • SqlSource.read()
      • SqlSource.schema()
  • Module contents
    • CsvSource
      • CsvSource.__init__()
      • CsvSource.read()
      • CsvSource.schema()
    • DataFrameSource
      • DataFrameSource.__init__()
      • DataFrameSource.read()
      • DataFrameSource.schema()
    • ParquetSource
      • ParquetSource.__init__()
      • ParquetSource.read()
      • ParquetSource.schema()
    • SqlSource
      • SqlSource.__init__()
      • SqlSource.read()
      • SqlSource.schema()

Submodules

tanat.store.source.base module

Abstract base class for all ingestion sources.

class tanat.store.source.base.AbstractSource

Bases: ABC, Registrable

Unified interface for all ingestion sources.

Subclasses implement read() and return a polars.LazyFrame. Registered by name so the resolver and builder can look them up by string.

Registration names: "csv", "parquet", "dataframe", "sql".

abstractmethod read() → LazyFrame

Read the source and return a lazy frame.

abstractmethod schema() → Schema

Return the column schema without reading the full dataset.

Implementations should be as cheap as possible - reading metadata only (file headers, SQL LIMIT 0 probe, in-memory schema, …).

Module contents

Data source package.

class tanat.store.source.AbstractSource

Bases: ABC, Registrable

Unified interface for all ingestion sources.

Subclasses implement read() and return a polars.LazyFrame. Registered by name so the resolver and builder can look them up by string.

Registration names: "csv", "parquet", "dataframe", "sql".

abstractmethod read() → LazyFrame

Read the source and return a lazy frame.

abstractmethod schema() → Schema

Return the column schema without reading the full dataset.

Implementations should be as cheap as possible - reading metadata only (file headers, SQL LIMIT 0 probe, in-memory schema, …).