mirror of
https://github.com/NousResearch/hermes-agent.git
synced 2026-07-15 14:22:43 +00:00
709da844b5
hermes send "MEDIA:/x.png This Caption" now arrives as one native captioned bubble instead of a separate text message followed by an uncaptioned bubble. Root cause: the standalone senders (hermes send / cron / send_message tool) stripped the MEDIA: tag, sent the remaining text as its own message, and called the media send with no caption -- even though hermes send's help advertises the captioned form and the bridges/adapters already support a caption. Signal already captioned correctly. - tools/send_message_tool.py: new _media_caption_split() chokepoint decides caption-vs-separate-body (single captionable non-voice file within the platform's message-length cap). Wired into the Telegram, WhatsApp and Discord dispatch paths. - Telegram/WhatsApp/Discord: when the single captioned file is missing, the caption text is delivered as a plain message so it is never silently lost. - Telegram caption send gets a MarkdownV2->plain parse fallback. - Tests: _media_caption_split unit tests + per-platform caption tests (ride, multi-file fallback, voice exclusion, over-limit fallback, missing-file text fallback); updated the 3 tests that asserted the old text-then-media split. Closes the gap reported against #58911 (the MEDIA_CAPTION directive PR); credit to @ferreiraesilva for surfacing the caption behavior.
296 lines
9.9 KiB
Python
296 lines
9.9 KiB
Python
"""WhatsApp media delivery for send_message (#19105).
|
|
|
|
Covers two layers:
|
|
|
|
* ``_bridge_media_type`` — extension/voice/force_document -> bridge mediaType.
|
|
* ``_standalone_send`` — text-first then per-file ``/send-media`` uploads,
|
|
media-only (skip ``/send``), and missing-file errors. The bridge HTTP calls
|
|
are mocked at the ``aiohttp.ClientSession`` boundary.
|
|
"""
|
|
|
|
import asyncio
|
|
import os
|
|
import tempfile
|
|
from types import SimpleNamespace
|
|
from unittest.mock import AsyncMock, MagicMock, patch
|
|
|
|
import pytest
|
|
|
|
from plugins.platforms.whatsapp.adapter import _bridge_media_type, _standalone_send
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# _bridge_media_type
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"path,is_voice,force_document,expected",
|
|
[
|
|
("a.png", False, False, "image"),
|
|
("a.JPG", False, False, "image"),
|
|
("a.jpeg", False, False, "image"),
|
|
("a.webp", False, False, "image"),
|
|
("a.gif", False, False, "image"),
|
|
("a.mp4", False, False, "video"),
|
|
("a.mov", False, False, "video"),
|
|
("a.webm", False, False, "video"),
|
|
("a.ogg", True, False, "audio"),
|
|
("a.opus", False, False, "audio"),
|
|
("a.mp3", False, False, "audio"),
|
|
("a.wav", False, False, "audio"),
|
|
("a.pdf", False, False, "document"),
|
|
("a.zip", False, False, "document"),
|
|
# force_document overrides everything
|
|
("a.png", False, True, "document"),
|
|
("a.mp4", False, True, "document"),
|
|
# is_voice wins over a video extension
|
|
("a.mp4", True, False, "audio"),
|
|
],
|
|
)
|
|
def test_bridge_media_type(path, is_voice, force_document, expected):
|
|
assert _bridge_media_type(path, is_voice, force_document) == expected
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# _standalone_send — bridge HTTP mocked
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def _resp(status, json_data=None, text_data=None):
|
|
r = AsyncMock()
|
|
r.status = status
|
|
r.json = AsyncMock(return_value=json_data or {})
|
|
r.text = AsyncMock(return_value=text_data or "")
|
|
return r
|
|
|
|
|
|
def _session_with(responses):
|
|
"""Build a mocked aiohttp.ClientSession that returns *responses* in order
|
|
and records every POST (url, json_payload)."""
|
|
calls = []
|
|
idx = [0]
|
|
|
|
def _post(url, **kwargs):
|
|
calls.append((url, kwargs.get("json")))
|
|
r = responses[idx[0]] if idx[0] < len(responses) else responses[-1]
|
|
idx[0] += 1
|
|
ctx = MagicMock()
|
|
ctx.__aenter__ = AsyncMock(return_value=r)
|
|
ctx.__aexit__ = AsyncMock(return_value=False)
|
|
return ctx
|
|
|
|
session = MagicMock()
|
|
session.post = MagicMock(side_effect=_post)
|
|
session_ctx = MagicMock()
|
|
session_ctx.__aenter__ = AsyncMock(return_value=session)
|
|
session_ctx.__aexit__ = AsyncMock(return_value=False)
|
|
return session_ctx, calls
|
|
|
|
|
|
def _pconfig():
|
|
return SimpleNamespace(token="", extra={"bridge_port": 3000})
|
|
|
|
|
|
def _tmpfile(suffix):
|
|
f = tempfile.NamedTemporaryFile(suffix=suffix, delete=False)
|
|
f.write(b"x")
|
|
f.close()
|
|
return f.name
|
|
|
|
|
|
def test_text_plus_mixed_media_routes_native_types():
|
|
img = _tmpfile(".png")
|
|
vid = _tmpfile(".mp4")
|
|
voice = _tmpfile(".ogg")
|
|
try:
|
|
session_ctx, calls = _session_with(
|
|
[
|
|
_resp(200, {"messageId": "t1"}),
|
|
_resp(200, {"messageId": "m1"}),
|
|
_resp(200, {"messageId": "m2"}),
|
|
_resp(200, {"messageId": "m3"}),
|
|
]
|
|
)
|
|
with patch("aiohttp.ClientSession", return_value=session_ctx):
|
|
res = asyncio.run(
|
|
_standalone_send(
|
|
_pconfig(),
|
|
"12345",
|
|
"hello",
|
|
media_files=[(img, False), (vid, False), (voice, True)],
|
|
)
|
|
)
|
|
assert res["success"] is True
|
|
# text first, then three media uploads in order
|
|
assert calls[0][0].endswith("/send")
|
|
assert calls[0][1]["message"] == "hello"
|
|
media_types = [c[1]["mediaType"] for c in calls if c[0].endswith("/send-media")]
|
|
assert media_types == ["image", "video", "audio"]
|
|
# chat id normalized to a WhatsApp JID
|
|
assert "@" in calls[0][1]["chatId"]
|
|
finally:
|
|
for p in (img, vid, voice):
|
|
os.unlink(p)
|
|
|
|
|
|
def test_media_only_skips_text_send():
|
|
img = _tmpfile(".jpg")
|
|
try:
|
|
session_ctx, calls = _session_with([_resp(200, {"messageId": "m1"})])
|
|
with patch("aiohttp.ClientSession", return_value=session_ctx):
|
|
res = asyncio.run(
|
|
_standalone_send(_pconfig(), "12345", "", media_files=[(img, False)])
|
|
)
|
|
assert res["success"] is True
|
|
assert all(c[0].endswith("/send-media") for c in calls)
|
|
finally:
|
|
os.unlink(img)
|
|
|
|
|
|
def test_force_document_sends_image_as_document():
|
|
img = _tmpfile(".png")
|
|
try:
|
|
session_ctx, calls = _session_with(
|
|
[_resp(200, {"messageId": "t1"}), _resp(200, {"messageId": "m1"})]
|
|
)
|
|
with patch("aiohttp.ClientSession", return_value=session_ctx):
|
|
res = asyncio.run(
|
|
_standalone_send(
|
|
_pconfig(),
|
|
"12345",
|
|
"doc",
|
|
media_files=[(img, False)],
|
|
force_document=True,
|
|
)
|
|
)
|
|
assert res["success"] is True
|
|
media_call = [c for c in calls if c[0].endswith("/send-media")][0]
|
|
assert media_call[1]["mediaType"] == "document"
|
|
assert media_call[1]["fileName"] == os.path.basename(img)
|
|
finally:
|
|
os.unlink(img)
|
|
|
|
|
|
def test_missing_media_file_errors():
|
|
session_ctx, _ = _session_with([_resp(200, {"messageId": "t1"})])
|
|
with patch("aiohttp.ClientSession", return_value=session_ctx):
|
|
res = asyncio.run(
|
|
_standalone_send(
|
|
_pconfig(),
|
|
"12345",
|
|
"hi",
|
|
media_files=[("/no/such/file.png", False)],
|
|
)
|
|
)
|
|
assert "error" in res
|
|
assert "not found" in res["error"]
|
|
|
|
|
|
def test_media_upload_error_propagates():
|
|
img = _tmpfile(".png")
|
|
try:
|
|
session_ctx, _ = _session_with(
|
|
[
|
|
_resp(200, {"messageId": "t1"}),
|
|
_resp(500, text_data="boom"),
|
|
]
|
|
)
|
|
with patch("aiohttp.ClientSession", return_value=session_ctx):
|
|
res = asyncio.run(
|
|
_standalone_send(
|
|
_pconfig(), "12345", "hi", media_files=[(img, False)]
|
|
)
|
|
)
|
|
assert "error" in res
|
|
assert "500" in res["error"]
|
|
finally:
|
|
os.unlink(img)
|
|
|
|
|
|
def test_text_only_unchanged_behavior():
|
|
session_ctx, calls = _session_with([_resp(200, {"messageId": "t1"})])
|
|
with patch("aiohttp.ClientSession", return_value=session_ctx):
|
|
res = asyncio.run(_standalone_send(_pconfig(), "12345", "just text"))
|
|
assert res == {
|
|
"success": True,
|
|
"platform": "whatsapp",
|
|
"chat_id": calls[0][1]["chatId"],
|
|
"message_id": "t1",
|
|
}
|
|
assert len(calls) == 1 and calls[0][0].endswith("/send")
|
|
|
|
|
|
def test_caption_rides_media_no_separate_text_send():
|
|
"""MEDIA:<path> caption -> single /send-media with caption, no /send."""
|
|
img = _tmpfile(".png")
|
|
try:
|
|
session_ctx, calls = _session_with([_resp(200, {"messageId": "m1"})])
|
|
with patch("aiohttp.ClientSession", return_value=session_ctx):
|
|
res = asyncio.run(
|
|
_standalone_send(
|
|
_pconfig(),
|
|
"12345",
|
|
"",
|
|
media_files=[(img, False)],
|
|
caption="2-bedroom floor plan",
|
|
)
|
|
)
|
|
assert res["success"] is True
|
|
# No separate /send — exactly one /send-media carrying the caption.
|
|
assert len(calls) == 1
|
|
assert calls[0][0].endswith("/send-media")
|
|
assert calls[0][1]["caption"] == "2-bedroom floor plan"
|
|
assert calls[0][1]["mediaType"] == "image"
|
|
finally:
|
|
os.unlink(img)
|
|
|
|
|
|
def test_caption_ignored_for_multi_file_send():
|
|
"""A caption never rides a multi-file send (association is ambiguous)."""
|
|
img = _tmpfile(".png")
|
|
img2 = _tmpfile(".jpg")
|
|
try:
|
|
session_ctx, calls = _session_with(
|
|
[_resp(200, {"messageId": "m1"}), _resp(200, {"messageId": "m2"})]
|
|
)
|
|
with patch("aiohttp.ClientSession", return_value=session_ctx):
|
|
res = asyncio.run(
|
|
_standalone_send(
|
|
_pconfig(),
|
|
"12345",
|
|
"",
|
|
media_files=[(img, False), (img2, False)],
|
|
caption="should be ignored",
|
|
)
|
|
)
|
|
assert res["success"] is True
|
|
media_calls = [c for c in calls if c[0].endswith("/send-media")]
|
|
assert len(media_calls) == 2
|
|
assert all("caption" not in c[1] for c in media_calls)
|
|
finally:
|
|
os.unlink(img)
|
|
os.unlink(img2)
|
|
|
|
|
|
def test_missing_captioned_file_falls_back_to_text():
|
|
"""If the single captioned file is missing, the caption is delivered as a
|
|
plain /send message rather than being silently lost (W1)."""
|
|
session_ctx, calls = _session_with([_resp(200, {"messageId": "t1"})])
|
|
with patch("aiohttp.ClientSession", return_value=session_ctx):
|
|
res = asyncio.run(
|
|
_standalone_send(
|
|
_pconfig(),
|
|
"12345",
|
|
"",
|
|
media_files=[("/no/such/file.png", False)],
|
|
caption="floor plan",
|
|
)
|
|
)
|
|
# The send still surfaces the missing-file error...
|
|
assert "error" in res
|
|
assert "not found" in res["error"]
|
|
# ...but the caption text was delivered on its own first.
|
|
assert len(calls) == 1
|
|
assert calls[0][0].endswith("/send")
|
|
assert calls[0][1]["message"] == "floor plan"
|