Links
AI / Agents
Meta
Requires: Python >=3.10
gdtest-skill-config
A streaming toolkit with enriched skill configuration.
Installation
pip install gdtest-skill-configAuto-generated skill enriched with gotchas, best_practices, decision_table, and extra_body from great-docs.yml. Tests that config overrides are injected into the generated skill.md body and rendered correctly on the Skills page with styled tables and code blocks.
Source files
gdtest_skill_config/
__init__.py
"""A streaming toolkit."""
__version__ = "0.2.0"
__all__ = ["Stream", "Sink", "batch"]
class Stream:
"""
A lazy stream reader.
Parameters
----------
source
A file path or file-like object.
chunk_size
Read chunk size in bytes.
timeout
Read timeout in seconds.
"""
def __init__(
self,
source: str,
chunk_size: int = 1024,
timeout: float | None = None,
):
self.source = source
self.chunk_size = chunk_size
self.timeout = timeout
def __enter__(self):
return self
def __exit__(self, *args):
self.close()
def read(self, n: int = -1) -> bytes:
"""
Read up to n bytes.
Parameters
----------
n
Number of bytes. -1 for all remaining.
Returns
-------
bytes
The data read.
"""
return b""
def close(self) -> None:
"""Close the stream and release resources."""
pass
def map(self, fn) -> "Stream":
"""
Apply a function to each element.
Parameters
----------
fn
Callable to apply.
Returns
-------
Stream
A new transformed stream.
"""
return self
def filter(self, predicate) -> "Stream":
"""
Filter elements by a predicate.
Parameters
----------
predicate
Callable returning bool.
Returns
-------
Stream
A new filtered stream.
"""
return self
class Sink:
"""
A buffered output sink.
Parameters
----------
dest
Destination file path or file-like object.
buffer_size
Write buffer size in bytes.
"""
def __init__(self, dest: str, buffer_size: int = 4096):
self.dest = dest
self.buffer_size = buffer_size
def write(self, data: bytes) -> int:
"""
Write data to the sink.
Parameters
----------
data
Bytes to write.
Returns
-------
int
Number of bytes written.
"""
return len(data)
def flush(self) -> None:
"""Flush the write buffer."""
pass
def batch(iterable, size: int = 10):
"""
Yield items from an iterable in fixed-size batches.
Parameters
----------
iterable
Input iterable.
size
Batch size.
Yields
------
list
A batch of items.
"""
batch_items = []
for item in iterable:
batch_items.append(item)
if len(batch_items) == size:
yield batch_items
batch_items = []
if batch_items:
yield batch_itemsREADME.md
# gdtest-skill-config A streaming toolkit with enriched skill configuration. ## Installation ```bash pip install gdtest-skill-config ```
great-docs.yml
skill:
gotchas:
- "Always close streams with `stream.close()` or use a context manager."
- "The `batch()` generator is lazy — it won't pull data until iterated."
- Default chunk size is 1024 bytes; increase for large payloads.
best_practices:
- "Use `with Stream(source) as s:` for automatic cleanup."
- "Prefer `batch()` over manual iteration for memory efficiency."
- "Set `timeout` explicitly in production to avoid hangs."
decision_table:
- if: Need to read data incrementally
then: stream = Stream(source)
- if: Need to process in fixed-size chunks
then: "for chunk in stream.batch(size=4096): ..."
- if: Need to transform each element
then: stream.map(fn) | stream.filter(pred)
- if: Need buffered output
then: "sink = Sink(dest, buffer_size=8192)"
extra_body: |-
## Performance notes
Throughput scales linearly up to ~10 concurrent streams. Beyond that, consider using `asyncio` or a thread pool.
## Compatibility
Works with file-like objects, sockets, and any object implementing the `Readable` protocol.