Skip to content

v1.2.0

Latest
Latest

Choose a tag to compare

View all tags
@sudoskys sudoskys released this 14 Jun 00:06
pypi_1.2.0
This tag was signed with the committer’s verified signature.
sudoskys Jasmine
GPG key ID: A1968AD61691308F
Verified
Learn about vigilant mode.
ad24560

What's New

Streaming draft support (ADR-002)

New DraftStream and EditStream classes for token-by-token LLM output via Telegram's draft API:

from telegramify_markdown.stream import DraftStream

async with DraftStream(
    send_draft=my_send_draft_fn,
    send_final=my_send_final_fn,
    mode="rich",
    interval=0.3,
    thinking_delay=0.5,
) as stream:
    async for token in llm_response:
        stream.feed(token)
  • DraftStream — for private chats via sendMessageDraft / sendRichMessageDraft (Bot API 9.3+)
  • EditStream — for group/channel chats via editMessageText (≥1.0s interval enforced)
  • StreamCore — protocol-agnostic base with throttling, keepalive, degraded mode, and concurrent send guard

Rich Message improvements

  • Oversized blocks (>32KB paragraph or code block) are now split into valid chunks instead of being emitted as-is with a warning
  • Empty richify() / telegramify_rich() input returns empty list instead of a single empty payload

Migration from v1.1.x

Fully backward-compatible. No API changes, no behavior regressions. Streaming is purely additive — import from telegramify_markdown.stream when needed.

Full Changelog

  • feat: add streaming draft support (ADR-002) — StreamCore, DraftStream, EditStream, 22 unit tests, live integration tests
  • fix: split oversized rich message blocks — paragraph, pre, and Markdown blocks exceeding 32KB are now split correctly
  • ci: add stream live test and unit test steps to CI
Assets 2
Loading