No public API for `COPY FROM STDIN` (text/binary stream ingestion)

[Dev-iL]

Summary

psqlpy exposes no public method for streaming data into PostgreSQL via the COPY FROM STDIN protocol. binary_copy_to_table exists on Connection but requires pre-encoded binary data in PostgreSQL's binary COPY format, which is not interchangeable with the text/CSV COPY FROM STDIN path used by every other Python driver.

Environment

• psqlpy: 0.11.12
• Python: 3.14.3
• PostgreSQL: 14.20

What other drivers provide

Driver	API
asyncpg	conn.copy_records_to_table(table, records=[...], columns=[...])
psycopg3	async with cur.copy("COPY t FROM STDIN") as copy: await copy.write(data)
psycopg2	cur.copy_from(file_obj, table, columns=[...])

COPY is the fastest bulk-load path in PostgreSQL (bypasses row-by-row parsing and WAL overhead for unlogged tables). Benchmarks against asyncpg's copy_records_to_table show it sustaining >1 M rows/sec on local connections.

What psqlpy currently provides

# Exists on Connection, but takes raw binary COPY format bytes:
conn.binary_copy_to_table(source, table_name, columns=None, schema_name=None)

There is no text/CSV COPY path, no record-list API, and no streaming write API analogous to psycopg3's async with cur.copy(...) as copy.

Expected / requested API

A high-level API that accepts Python objects (or at least raw text/bytes in PostgreSQL text COPY format) without requiring callers to produce binary COPY wire format manually:

# Option A: record list (like asyncpg)
await conn.copy_records_to_table(
    "my_table",
    records=[(1, "a"), (2, "b")],
    columns=["id", "name"],
)

# Option B: streaming writer (like psycopg3)
async with conn.copy("COPY my_table FROM STDIN") as writer:
    await writer.write("1\ta\n2\tb\n")

Impact

Any application that bulk-loads data — ETL pipelines, data import scripts, benchmarks — cannot use psqlpy for this workload without a lossy workaround (falling back to execute_many, which is significantly slower; see related issue on execute_many performance).