feat: add SearX search plugin and alert backend
Add standalone !searx command for on-demand SearXNG search (top 3 results). Add SearX as a third backend (sx) to the alert plugin for keyword monitoring. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
@@ -341,12 +341,20 @@ No API credentials needed (uses public GQL endpoint).
|
||||
!alert check <name> # Force-poll now
|
||||
```
|
||||
|
||||
Searches keywords across YouTube (InnerTube) and Twitch (GQL) simultaneously.
|
||||
Searches keywords across YouTube (InnerTube), Twitch (GQL), and SearXNG simultaneously.
|
||||
Names: lowercase alphanumeric + hyphens, 1-20 chars. Keywords: 1-100 chars.
|
||||
Max 20 alerts/channel. Polls every 5min. Max 5 announcements per platform per cycle.
|
||||
Format: `[name/yt] Title -- URL` or `[name/tw] Title -- URL`.
|
||||
Format: `[name/yt] Title -- URL`, `[name/tw] Title -- URL`, or `[name/sx] Title -- URL`.
|
||||
No API credentials needed. Persists across restarts.
|
||||
|
||||
## SearX
|
||||
|
||||
```
|
||||
!searx <query> # Search SearXNG
|
||||
```
|
||||
|
||||
Shows top 3 results as `Title -- URL`. Channel only. Max query length: 200 chars.
|
||||
|
||||
## Plugin Template
|
||||
|
||||
```python
|
||||
|
||||
@@ -122,6 +122,7 @@ format = "text" # Log format: "text" (default) or "json"
|
||||
| `!username <user>` | Check username across ~25 services |
|
||||
| `!username <user> <service>` | Check single service |
|
||||
| `!username list` | Show available services by category |
|
||||
| `!searx <query>` | Search SearXNG and show top results |
|
||||
|
||||
### Command Shorthand
|
||||
|
||||
@@ -630,11 +631,34 @@ Polling and announcements:
|
||||
- `list` shows live/error status indicators next to each streamer
|
||||
- `check` forces an immediate poll and reports current status
|
||||
|
||||
### `!searx` -- SearXNG Web Search
|
||||
|
||||
Search the local SearXNG instance and display top results.
|
||||
|
||||
```
|
||||
!searx <query...> Search SearXNG and show top results
|
||||
```
|
||||
|
||||
- Open to all users, channel only (no PM)
|
||||
- Query is everything after `!searx`
|
||||
- Shows top 3 results as `Title -- URL`
|
||||
- Titles truncated to 80 characters
|
||||
- Query limited to 200 characters
|
||||
|
||||
Output format:
|
||||
|
||||
```
|
||||
Title One -- https://example.com/page1
|
||||
Title Two -- https://example.com/page2
|
||||
Title Three -- https://example.com/page3
|
||||
```
|
||||
|
||||
### `!alert` -- Keyword Alert Subscriptions
|
||||
|
||||
Search keywords across multiple platforms (YouTube, Twitch) and announce new
|
||||
results. Unlike `!rss`/`!yt`/`!twitch` which follow specific channels/feeds,
|
||||
`!alert` searches keywords across all supported platforms simultaneously.
|
||||
Search keywords across multiple platforms (YouTube, Twitch, SearXNG) and
|
||||
announce new results. Unlike `!rss`/`!yt`/`!twitch` which follow specific
|
||||
channels/feeds, `!alert` searches keywords across all supported platforms
|
||||
simultaneously.
|
||||
|
||||
```
|
||||
!alert add <name> <keyword...> Add keyword alert (admin)
|
||||
@@ -654,12 +678,14 @@ Platforms searched:
|
||||
|
||||
- **YouTube** -- InnerTube search API (no auth required)
|
||||
- **Twitch** -- Public GQL endpoint: live streams and VODs (no auth required)
|
||||
- **SearXNG** -- Local SearXNG instance (no auth required)
|
||||
|
||||
Polling and announcements:
|
||||
|
||||
- Alerts are polled every 5 minutes by default
|
||||
- On `add`, existing results are recorded without announcing (prevents flood)
|
||||
- New results announced as `[name/yt] Title -- URL` or `[name/tw] Title -- URL`
|
||||
- New results announced as `[name/yt] Title -- URL`, `[name/tw] Title -- URL`,
|
||||
or `[name/sx] Title -- URL`
|
||||
- Maximum 5 items announced per platform per poll; excess shown as `... and N more`
|
||||
- Titles are truncated to 80 characters
|
||||
- Each platform maintains its own seen list (capped at 200 per platform)
|
||||
|
||||
@@ -26,6 +26,7 @@ _YT_SEARCH_URL = "https://www.youtube.com/youtubei/v1/search"
|
||||
_YT_CLIENT_VERSION = "2.20250101.00.00"
|
||||
_GQL_URL = "https://gql.twitch.tv/gql"
|
||||
_GQL_CLIENT_ID = "kimne78kx3ncx6brgo4mv6wki5h1ko"
|
||||
_SEARX_URL = "http://192.168.122.119:3000/search"
|
||||
|
||||
# -- Module-level tracking ---------------------------------------------------
|
||||
|
||||
@@ -193,11 +194,40 @@ def _search_twitch(keyword: str) -> list[dict]:
|
||||
return results
|
||||
|
||||
|
||||
# -- SearXNG search (blocking) ----------------------------------------------
|
||||
|
||||
def _search_searx(keyword: str) -> list[dict]:
|
||||
"""Search SearXNG. Blocking."""
|
||||
import urllib.parse
|
||||
|
||||
params = urllib.parse.urlencode({"q": keyword, "format": "json"})
|
||||
url = f"{_SEARX_URL}?{params}"
|
||||
|
||||
req = urllib.request.Request(url, method="GET")
|
||||
resp = urllib.request.urlopen(req, timeout=_FETCH_TIMEOUT)
|
||||
raw = resp.read()
|
||||
resp.close()
|
||||
|
||||
data = json.loads(raw)
|
||||
results: list[dict] = []
|
||||
for item in data.get("results", []):
|
||||
item_url = item.get("url", "")
|
||||
title = item.get("title", "")
|
||||
results.append({
|
||||
"id": item_url,
|
||||
"title": title,
|
||||
"url": item_url,
|
||||
"extra": "",
|
||||
})
|
||||
return results
|
||||
|
||||
|
||||
# -- Backend registry -------------------------------------------------------
|
||||
|
||||
_BACKENDS: dict[str, callable] = {
|
||||
"yt": _search_youtube,
|
||||
"tw": _search_twitch,
|
||||
"sx": _search_searx,
|
||||
}
|
||||
|
||||
|
||||
|
||||
94
plugins/searx.py
Normal file
94
plugins/searx.py
Normal file
@@ -0,0 +1,94 @@
|
||||
"""Plugin: SearXNG web search."""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import json
|
||||
import urllib.parse
|
||||
import urllib.request
|
||||
|
||||
from derp.plugin import command
|
||||
|
||||
# -- Constants ---------------------------------------------------------------
|
||||
|
||||
_SEARX_URL = "http://192.168.122.119:3000/search"
|
||||
_FETCH_TIMEOUT = 10
|
||||
_MAX_RESULTS = 3
|
||||
_MAX_TITLE_LEN = 80
|
||||
_MAX_QUERY_LEN = 200
|
||||
|
||||
|
||||
# -- Pure helpers ------------------------------------------------------------
|
||||
|
||||
def _truncate(text: str, max_len: int = _MAX_TITLE_LEN) -> str:
|
||||
"""Truncate text with ellipsis if needed."""
|
||||
if len(text) <= max_len:
|
||||
return text
|
||||
return text[: max_len - 3].rstrip() + "..."
|
||||
|
||||
|
||||
# -- SearXNG search (blocking) ----------------------------------------------
|
||||
|
||||
def _search(query: str) -> list[dict]:
|
||||
"""Search SearXNG. Blocking.
|
||||
|
||||
Returns list of dicts with keys: title, url, snippet.
|
||||
Raises on HTTP or parse errors.
|
||||
"""
|
||||
params = urllib.parse.urlencode({"q": query, "format": "json"})
|
||||
url = f"{_SEARX_URL}?{params}"
|
||||
|
||||
req = urllib.request.Request(url, method="GET")
|
||||
resp = urllib.request.urlopen(req, timeout=_FETCH_TIMEOUT)
|
||||
raw = resp.read()
|
||||
resp.close()
|
||||
|
||||
data = json.loads(raw)
|
||||
results: list[dict] = []
|
||||
for item in data.get("results", []):
|
||||
results.append({
|
||||
"title": item.get("title", ""),
|
||||
"url": item.get("url", ""),
|
||||
"snippet": item.get("content", item.get("snippet", "")),
|
||||
})
|
||||
return results
|
||||
|
||||
|
||||
# -- Command handler ---------------------------------------------------------
|
||||
|
||||
@command("searx", help="Search: !searx <query>")
|
||||
async def cmd_searx(bot, message):
|
||||
"""Search SearXNG and show top results.
|
||||
|
||||
Usage: !searx <query...>
|
||||
"""
|
||||
if not message.is_channel:
|
||||
await bot.reply(message, "Use this command in a channel")
|
||||
return
|
||||
|
||||
parts = message.text.split(None, 1)
|
||||
if len(parts) < 2 or not parts[1].strip():
|
||||
await bot.reply(message, "Usage: !searx <query>")
|
||||
return
|
||||
|
||||
query = parts[1].strip()
|
||||
if len(query) > _MAX_QUERY_LEN:
|
||||
await bot.reply(message, f"Query too long (max {_MAX_QUERY_LEN} chars)")
|
||||
return
|
||||
|
||||
import asyncio
|
||||
loop = asyncio.get_running_loop()
|
||||
|
||||
try:
|
||||
results = await loop.run_in_executor(None, _search, query)
|
||||
except Exception as exc:
|
||||
await bot.reply(message, f"Search failed: {exc}")
|
||||
return
|
||||
|
||||
if not results:
|
||||
await bot.reply(message, f"No results for: {query}")
|
||||
return
|
||||
|
||||
for item in results[:_MAX_RESULTS]:
|
||||
title = _truncate(item["title"]) if item["title"] else "(no title)"
|
||||
url = item["url"]
|
||||
await bot.reply(message, f"{title} -- {url}")
|
||||
@@ -28,6 +28,7 @@ from plugins.alert import ( # noqa: E402
|
||||
_pollers,
|
||||
_restore,
|
||||
_save,
|
||||
_search_searx,
|
||||
_search_twitch,
|
||||
_search_youtube,
|
||||
_start_poller,
|
||||
@@ -117,6 +118,15 @@ GQL_RESPONSE = {
|
||||
},
|
||||
}
|
||||
|
||||
# SearXNG search response
|
||||
SEARX_RESPONSE = {
|
||||
"results": [
|
||||
{"title": "SearX Result 1", "url": "https://example.com/sx1", "content": "Snippet 1"},
|
||||
{"title": "SearX Result 2", "url": "https://example.com/sx2", "content": "Snippet 2"},
|
||||
{"title": "SearX Result 3", "url": "https://example.com/sx3", "content": "Snippet 3"},
|
||||
],
|
||||
}
|
||||
|
||||
|
||||
# -- Helpers -----------------------------------------------------------------
|
||||
|
||||
@@ -218,7 +228,22 @@ def _fake_tw_error(keyword):
|
||||
raise ConnectionError("Twitch down")
|
||||
|
||||
|
||||
_FAKE_BACKENDS = {"yt": _fake_yt, "tw": _fake_tw}
|
||||
def _fake_sx(keyword):
|
||||
"""Fake SearX backend returning two results."""
|
||||
return [
|
||||
{"id": "https://example.com/sx1", "title": "SX Result 1",
|
||||
"url": "https://example.com/sx1", "extra": ""},
|
||||
{"id": "https://example.com/sx2", "title": "SX Result 2",
|
||||
"url": "https://example.com/sx2", "extra": ""},
|
||||
]
|
||||
|
||||
|
||||
def _fake_sx_error(keyword):
|
||||
"""Fake SearX backend that raises."""
|
||||
raise ConnectionError("SearX down")
|
||||
|
||||
|
||||
_FAKE_BACKENDS = {"yt": _fake_yt, "tw": _fake_tw, "sx": _fake_sx}
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
@@ -512,6 +537,7 @@ class TestCmdAlertAdd:
|
||||
assert "minecraft speedrun" in bot.replied[0]
|
||||
assert "2 yt" in bot.replied[0]
|
||||
assert "2 tw" in bot.replied[0]
|
||||
assert "2 sx" in bot.replied[0]
|
||||
data = _load(bot, "#test:mc-speed")
|
||||
assert data is not None
|
||||
assert data["name"] == "mc-speed"
|
||||
@@ -519,6 +545,7 @@ class TestCmdAlertAdd:
|
||||
assert data["channel"] == "#test"
|
||||
assert len(data["seen"]["yt"]) == 2
|
||||
assert len(data["seen"]["tw"]) == 2
|
||||
assert len(data["seen"]["sx"]) == 2
|
||||
assert "#test:mc-speed" in _pollers
|
||||
_stop_poller("#test:mc-speed")
|
||||
await asyncio.sleep(0)
|
||||
@@ -590,7 +617,7 @@ class TestCmdAlertAdd:
|
||||
"""If a backend fails during seeding, seen list is empty for that backend."""
|
||||
_clear()
|
||||
bot = _FakeBot(admin=True)
|
||||
backends = {"yt": _fake_yt, "tw": _fake_tw_error}
|
||||
backends = {"yt": _fake_yt, "tw": _fake_tw_error, "sx": _fake_sx}
|
||||
|
||||
async def inner():
|
||||
with patch.object(_mod, "_BACKENDS", backends):
|
||||
@@ -600,6 +627,7 @@ class TestCmdAlertAdd:
|
||||
assert data is not None
|
||||
assert len(data["seen"]["yt"]) == 2
|
||||
assert len(data["seen"]["tw"]) == 0
|
||||
assert len(data["seen"]["sx"]) == 2
|
||||
_stop_poller("#test:partial")
|
||||
await asyncio.sleep(0)
|
||||
|
||||
@@ -722,7 +750,10 @@ class TestCmdAlertCheck:
|
||||
bot = _FakeBot()
|
||||
data = {
|
||||
"keyword": "test", "name": "chk", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": ["yt1", "yt2"], "tw": ["stream:tw1", "vod:tw2"]},
|
||||
"interval": 300, "seen": {
|
||||
"yt": ["yt1", "yt2"], "tw": ["stream:tw1", "vod:tw2"],
|
||||
"sx": ["https://example.com/sx1", "https://example.com/sx2"],
|
||||
},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
_save(bot, "#test:chk", data)
|
||||
@@ -751,11 +782,11 @@ class TestCmdAlertCheck:
|
||||
bot = _FakeBot()
|
||||
data = {
|
||||
"keyword": "test", "name": "errchk", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": [], "tw": []},
|
||||
"interval": 300, "seen": {"yt": [], "tw": [], "sx": []},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
_save(bot, "#test:errchk", data)
|
||||
backends = {"yt": _fake_yt_error, "tw": _fake_tw}
|
||||
backends = {"yt": _fake_yt_error, "tw": _fake_tw, "sx": _fake_sx}
|
||||
|
||||
async def inner():
|
||||
with patch.object(_mod, "_BACKENDS", backends):
|
||||
@@ -769,7 +800,7 @@ class TestCmdAlertCheck:
|
||||
bot = _FakeBot()
|
||||
data = {
|
||||
"keyword": "test", "name": "news", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": ["yt1"], "tw": []},
|
||||
"interval": 300, "seen": {"yt": ["yt1"], "tw": [], "sx": []},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
_save(bot, "#test:news", data)
|
||||
@@ -777,12 +808,14 @@ class TestCmdAlertCheck:
|
||||
async def inner():
|
||||
with patch.object(_mod, "_BACKENDS", _FAKE_BACKENDS):
|
||||
await cmd_alert(bot, _msg("!alert check news"))
|
||||
# yt2 is new for yt, both tw results are new
|
||||
# yt2 is new for yt, both tw and sx results are new
|
||||
announcements = [s for t, s in bot.sent if t == "#test"]
|
||||
yt_msgs = [m for m in announcements if "/yt]" in m]
|
||||
tw_msgs = [m for m in announcements if "/tw]" in m]
|
||||
sx_msgs = [m for m in announcements if "/sx]" in m]
|
||||
assert len(yt_msgs) == 1 # yt2 only
|
||||
assert len(tw_msgs) == 2 # both tw results
|
||||
assert len(sx_msgs) == 2 # both sx results
|
||||
|
||||
asyncio.run(inner())
|
||||
|
||||
@@ -803,7 +836,7 @@ class TestPollOnce:
|
||||
bot = _FakeBot()
|
||||
data = {
|
||||
"keyword": "test", "name": "poll", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": [], "tw": []},
|
||||
"interval": 300, "seen": {"yt": [], "tw": [], "sx": []},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
key = "#test:poll"
|
||||
@@ -814,9 +847,10 @@ class TestPollOnce:
|
||||
with patch.object(_mod, "_BACKENDS", _FAKE_BACKENDS):
|
||||
await _poll_once(bot, key, announce=True)
|
||||
messages = [s for t, s in bot.sent if t == "#test"]
|
||||
assert len(messages) == 4 # 2 yt + 2 tw
|
||||
assert len(messages) == 6 # 2 yt + 2 tw + 2 sx
|
||||
assert "[poll/yt]" in messages[0]
|
||||
assert "[poll/tw]" in messages[2]
|
||||
assert "[poll/sx]" in messages[4]
|
||||
|
||||
asyncio.run(inner())
|
||||
|
||||
@@ -826,7 +860,10 @@ class TestPollOnce:
|
||||
data = {
|
||||
"keyword": "test", "name": "dedup", "channel": "#test",
|
||||
"interval": 300,
|
||||
"seen": {"yt": ["yt1", "yt2"], "tw": ["stream:tw1", "vod:tw2"]},
|
||||
"seen": {
|
||||
"yt": ["yt1", "yt2"], "tw": ["stream:tw1", "vod:tw2"],
|
||||
"sx": ["https://example.com/sx1", "https://example.com/sx2"],
|
||||
},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
key = "#test:dedup"
|
||||
@@ -854,7 +891,7 @@ class TestPollOnce:
|
||||
|
||||
data = {
|
||||
"keyword": "test", "name": "many", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": [], "tw": []},
|
||||
"interval": 300, "seen": {"yt": [], "tw": [], "sx": []},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
key = "#test:many"
|
||||
@@ -876,20 +913,22 @@ class TestPollOnce:
|
||||
bot = _FakeBot()
|
||||
data = {
|
||||
"keyword": "test", "name": "partial", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": [], "tw": []},
|
||||
"interval": 300, "seen": {"yt": [], "tw": [], "sx": []},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
key = "#test:partial"
|
||||
_save(bot, key, data)
|
||||
_subscriptions[key] = data
|
||||
backends = {"yt": _fake_yt_error, "tw": _fake_tw}
|
||||
backends = {"yt": _fake_yt_error, "tw": _fake_tw, "sx": _fake_sx}
|
||||
|
||||
async def inner():
|
||||
with patch.object(_mod, "_BACKENDS", backends):
|
||||
await _poll_once(bot, key, announce=True)
|
||||
# Twitch results should still be announced
|
||||
# Twitch and SearX results should still be announced
|
||||
tw_msgs = [s for t, s in bot.sent if t == "#test" and "/tw]" in s]
|
||||
sx_msgs = [s for t, s in bot.sent if t == "#test" and "/sx]" in s]
|
||||
assert len(tw_msgs) == 2
|
||||
assert len(sx_msgs) == 2
|
||||
# Error counter should be incremented
|
||||
assert _errors[key] == 1
|
||||
updated = _load(bot, key)
|
||||
@@ -902,7 +941,7 @@ class TestPollOnce:
|
||||
bot = _FakeBot()
|
||||
data = {
|
||||
"keyword": "test", "name": "quiet", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": [], "tw": []},
|
||||
"interval": 300, "seen": {"yt": [], "tw": [], "sx": []},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
key = "#test:quiet"
|
||||
@@ -916,6 +955,7 @@ class TestPollOnce:
|
||||
updated = _load(bot, key)
|
||||
assert len(updated["seen"]["yt"]) == 2
|
||||
assert len(updated["seen"]["tw"]) == 2
|
||||
assert len(updated["seen"]["sx"]) == 2
|
||||
|
||||
asyncio.run(inner())
|
||||
|
||||
@@ -932,7 +972,7 @@ class TestPollOnce:
|
||||
|
||||
data = {
|
||||
"keyword": "test", "name": "cap", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": [], "tw": []},
|
||||
"interval": 300, "seen": {"yt": [], "tw": [], "sx": []},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
key = "#test:cap"
|
||||
@@ -954,13 +994,13 @@ class TestPollOnce:
|
||||
bot = _FakeBot()
|
||||
data = {
|
||||
"keyword": "test", "name": "allerr", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": [], "tw": []},
|
||||
"interval": 300, "seen": {"yt": [], "tw": [], "sx": []},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
key = "#test:allerr"
|
||||
_save(bot, key, data)
|
||||
_subscriptions[key] = data
|
||||
backends = {"yt": _fake_yt_error, "tw": _fake_tw_error}
|
||||
backends = {"yt": _fake_yt_error, "tw": _fake_tw_error, "sx": _fake_sx_error}
|
||||
|
||||
async def inner():
|
||||
with patch.object(_mod, "_BACKENDS", backends):
|
||||
@@ -975,7 +1015,10 @@ class TestPollOnce:
|
||||
bot = _FakeBot()
|
||||
data = {
|
||||
"keyword": "test", "name": "clrerr", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": ["yt1", "yt2"], "tw": ["stream:tw1", "vod:tw2"]},
|
||||
"interval": 300, "seen": {
|
||||
"yt": ["yt1", "yt2"], "tw": ["stream:tw1", "vod:tw2"],
|
||||
"sx": ["https://example.com/sx1", "https://example.com/sx2"],
|
||||
},
|
||||
"last_poll": "", "last_error": "old error",
|
||||
}
|
||||
key = "#test:clrerr"
|
||||
@@ -1003,7 +1046,7 @@ class TestRestore:
|
||||
bot = _FakeBot()
|
||||
data = {
|
||||
"keyword": "test", "name": "restored", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": [], "tw": []},
|
||||
"interval": 300, "seen": {"yt": [], "tw": [], "sx": []},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
_save(bot, "#test:restored", data)
|
||||
@@ -1023,7 +1066,7 @@ class TestRestore:
|
||||
bot = _FakeBot()
|
||||
data = {
|
||||
"keyword": "test", "name": "active", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": [], "tw": []},
|
||||
"interval": 300, "seen": {"yt": [], "tw": [], "sx": []},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
_save(bot, "#test:active", data)
|
||||
@@ -1043,7 +1086,7 @@ class TestRestore:
|
||||
bot = _FakeBot()
|
||||
data = {
|
||||
"keyword": "test", "name": "done", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": [], "tw": []},
|
||||
"interval": 300, "seen": {"yt": [], "tw": [], "sx": []},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
_save(bot, "#test:done", data)
|
||||
@@ -1077,7 +1120,7 @@ class TestRestore:
|
||||
bot = _FakeBot()
|
||||
data = {
|
||||
"keyword": "test", "name": "conn", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": [], "tw": []},
|
||||
"interval": 300, "seen": {"yt": [], "tw": [], "sx": []},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
_save(bot, "#test:conn", data)
|
||||
@@ -1102,7 +1145,7 @@ class TestPollerManagement:
|
||||
bot = _FakeBot()
|
||||
data = {
|
||||
"keyword": "test", "name": "mgmt", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": [], "tw": []},
|
||||
"interval": 300, "seen": {"yt": [], "tw": [], "sx": []},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
key = "#test:mgmt"
|
||||
@@ -1125,7 +1168,7 @@ class TestPollerManagement:
|
||||
bot = _FakeBot()
|
||||
data = {
|
||||
"keyword": "test", "name": "idem", "channel": "#test",
|
||||
"interval": 300, "seen": {"yt": [], "tw": []},
|
||||
"interval": 300, "seen": {"yt": [], "tw": [], "sx": []},
|
||||
"last_poll": "", "last_error": "",
|
||||
}
|
||||
key = "#test:idem"
|
||||
@@ -1163,3 +1206,44 @@ class TestCmdAlertUsage:
|
||||
bot = _FakeBot()
|
||||
asyncio.run(cmd_alert(bot, _msg("!alert foobar")))
|
||||
assert "Usage:" in bot.replied[0]
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# TestSearchSearx
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
class TestSearchSearx:
|
||||
def test_parses_response(self):
|
||||
class FakeResp:
|
||||
def read(self):
|
||||
return json.dumps(SEARX_RESPONSE).encode()
|
||||
def close(self):
|
||||
pass
|
||||
|
||||
with patch("urllib.request.urlopen", return_value=FakeResp()):
|
||||
results = _search_searx("test query")
|
||||
assert len(results) == 3
|
||||
assert results[0]["id"] == "https://example.com/sx1"
|
||||
assert results[0]["title"] == "SearX Result 1"
|
||||
assert results[0]["url"] == "https://example.com/sx1"
|
||||
assert results[0]["extra"] == ""
|
||||
|
||||
def test_empty_results(self):
|
||||
empty = {"results": []}
|
||||
|
||||
class FakeResp:
|
||||
def read(self):
|
||||
return json.dumps(empty).encode()
|
||||
def close(self):
|
||||
pass
|
||||
|
||||
with patch("urllib.request.urlopen", return_value=FakeResp()):
|
||||
results = _search_searx("nothing")
|
||||
assert results == []
|
||||
|
||||
def test_http_error_propagates(self):
|
||||
import pytest
|
||||
|
||||
with patch("urllib.request.urlopen", side_effect=ConnectionError("fail")):
|
||||
with pytest.raises(ConnectionError):
|
||||
_search_searx("test")
|
||||
|
||||
243
tests/test_searx.py
Normal file
243
tests/test_searx.py
Normal file
@@ -0,0 +1,243 @@
|
||||
"""Tests for the SearXNG search plugin."""
|
||||
|
||||
import asyncio
|
||||
import importlib.util
|
||||
import json
|
||||
import sys
|
||||
from pathlib import Path
|
||||
from unittest.mock import patch
|
||||
|
||||
from derp.irc import Message
|
||||
|
||||
# plugins/ is not a Python package -- load the module from file path
|
||||
_spec = importlib.util.spec_from_file_location(
|
||||
"plugins.searx", Path(__file__).resolve().parent.parent / "plugins" / "searx.py",
|
||||
)
|
||||
_mod = importlib.util.module_from_spec(_spec)
|
||||
sys.modules[_spec.name] = _mod
|
||||
_spec.loader.exec_module(_mod)
|
||||
|
||||
from plugins.searx import ( # noqa: E402
|
||||
_MAX_QUERY_LEN,
|
||||
_MAX_RESULTS,
|
||||
_search,
|
||||
_truncate,
|
||||
cmd_searx,
|
||||
)
|
||||
|
||||
# -- Fixtures ----------------------------------------------------------------
|
||||
|
||||
SEARX_RESPONSE = {
|
||||
"results": [
|
||||
{"title": "Result One", "url": "https://example.com/1", "content": "First snippet"},
|
||||
{"title": "Result Two", "url": "https://example.com/2", "content": "Second snippet"},
|
||||
{"title": "Result Three", "url": "https://example.com/3", "content": "Third snippet"},
|
||||
{"title": "Result Four", "url": "https://example.com/4", "content": "Fourth snippet"},
|
||||
{"title": "Result Five", "url": "https://example.com/5", "content": "Fifth snippet"},
|
||||
],
|
||||
}
|
||||
|
||||
SEARX_EMPTY = {"results": []}
|
||||
|
||||
|
||||
# -- Helpers -----------------------------------------------------------------
|
||||
|
||||
class _FakeBot:
|
||||
"""Minimal bot stand-in that captures sent/replied messages."""
|
||||
|
||||
def __init__(self):
|
||||
self.sent: list[tuple[str, str]] = []
|
||||
self.replied: list[str] = []
|
||||
|
||||
async def send(self, target: str, text: str) -> None:
|
||||
self.sent.append((target, text))
|
||||
|
||||
async def reply(self, message, text: str) -> None:
|
||||
self.replied.append(text)
|
||||
|
||||
|
||||
def _msg(text: str, nick: str = "alice", target: str = "#test") -> Message:
|
||||
"""Create a channel PRIVMSG."""
|
||||
return Message(
|
||||
raw="", prefix=f"{nick}!~{nick}@host", nick=nick,
|
||||
command="PRIVMSG", params=[target, text], tags={},
|
||||
)
|
||||
|
||||
|
||||
def _pm(text: str, nick: str = "alice") -> Message:
|
||||
"""Create a private PRIVMSG."""
|
||||
return Message(
|
||||
raw="", prefix=f"{nick}!~{nick}@host", nick=nick,
|
||||
command="PRIVMSG", params=["botname", text], tags={},
|
||||
)
|
||||
|
||||
|
||||
class _FakeResp:
|
||||
"""Fake HTTP response returning preset JSON."""
|
||||
|
||||
def __init__(self, data):
|
||||
self._data = data
|
||||
|
||||
def read(self):
|
||||
return json.dumps(self._data).encode()
|
||||
|
||||
def close(self):
|
||||
pass
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# TestTruncate
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
class TestTruncate:
|
||||
def test_short_text_unchanged(self):
|
||||
assert _truncate("hello", 80) == "hello"
|
||||
|
||||
def test_exact_length_unchanged(self):
|
||||
text = "a" * 80
|
||||
assert _truncate(text, 80) == text
|
||||
|
||||
def test_long_text_truncated(self):
|
||||
text = "a" * 100
|
||||
result = _truncate(text, 80)
|
||||
assert len(result) == 80
|
||||
assert result.endswith("...")
|
||||
|
||||
def test_default_max_length(self):
|
||||
text = "a" * 100
|
||||
result = _truncate(text)
|
||||
assert len(result) == 80
|
||||
|
||||
def test_trailing_space_stripped(self):
|
||||
text = "word " * 20
|
||||
result = _truncate(text, 20)
|
||||
assert not result.endswith(" ...")
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# TestSearch
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
class TestSearch:
|
||||
def test_success(self):
|
||||
with patch("urllib.request.urlopen", return_value=_FakeResp(SEARX_RESPONSE)):
|
||||
results = _search("test query")
|
||||
assert len(results) == 5
|
||||
assert results[0]["title"] == "Result One"
|
||||
assert results[0]["url"] == "https://example.com/1"
|
||||
assert results[0]["snippet"] == "First snippet"
|
||||
|
||||
def test_empty_results(self):
|
||||
with patch("urllib.request.urlopen", return_value=_FakeResp(SEARX_EMPTY)):
|
||||
results = _search("nothing")
|
||||
assert results == []
|
||||
|
||||
def test_http_error_propagates(self):
|
||||
import pytest
|
||||
|
||||
with patch("urllib.request.urlopen", side_effect=ConnectionError("down")):
|
||||
with pytest.raises(ConnectionError):
|
||||
_search("test")
|
||||
|
||||
def test_snippet_fallback(self):
|
||||
"""Falls back to 'snippet' key when 'content' is absent."""
|
||||
data = {"results": [
|
||||
{"title": "T", "url": "http://x.com", "snippet": "fallback"},
|
||||
]}
|
||||
with patch("urllib.request.urlopen", return_value=_FakeResp(data)):
|
||||
results = _search("test")
|
||||
assert results[0]["snippet"] == "fallback"
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# TestCmdSearx
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
class TestCmdSearx:
|
||||
def test_results_displayed(self):
|
||||
bot = _FakeBot()
|
||||
|
||||
async def inner():
|
||||
with patch("urllib.request.urlopen", return_value=_FakeResp(SEARX_RESPONSE)):
|
||||
await cmd_searx(bot, _msg("!searx test query"))
|
||||
assert len(bot.replied) == _MAX_RESULTS
|
||||
assert "Result One" in bot.replied[0]
|
||||
assert "https://example.com/1" in bot.replied[0]
|
||||
assert " -- " in bot.replied[0]
|
||||
|
||||
asyncio.run(inner())
|
||||
|
||||
def test_no_results(self):
|
||||
bot = _FakeBot()
|
||||
|
||||
async def inner():
|
||||
with patch("urllib.request.urlopen", return_value=_FakeResp(SEARX_EMPTY)):
|
||||
await cmd_searx(bot, _msg("!searx nothing"))
|
||||
assert len(bot.replied) == 1
|
||||
assert "No results for:" in bot.replied[0]
|
||||
|
||||
asyncio.run(inner())
|
||||
|
||||
def test_error_handling(self):
|
||||
bot = _FakeBot()
|
||||
|
||||
async def inner():
|
||||
with patch("urllib.request.urlopen", side_effect=ConnectionError("fail")):
|
||||
await cmd_searx(bot, _msg("!searx broken"))
|
||||
assert len(bot.replied) == 1
|
||||
assert "Search failed:" in bot.replied[0]
|
||||
|
||||
asyncio.run(inner())
|
||||
|
||||
def test_pm_rejected(self):
|
||||
bot = _FakeBot()
|
||||
asyncio.run(cmd_searx(bot, _pm("!searx test")))
|
||||
assert "Use this command in a channel" in bot.replied[0]
|
||||
|
||||
def test_no_query(self):
|
||||
bot = _FakeBot()
|
||||
asyncio.run(cmd_searx(bot, _msg("!searx")))
|
||||
assert "Usage:" in bot.replied[0]
|
||||
|
||||
def test_empty_query(self):
|
||||
bot = _FakeBot()
|
||||
asyncio.run(cmd_searx(bot, _msg("!searx ")))
|
||||
assert "Usage:" in bot.replied[0]
|
||||
|
||||
def test_query_too_long(self):
|
||||
bot = _FakeBot()
|
||||
long_query = "x" * (_MAX_QUERY_LEN + 1)
|
||||
asyncio.run(cmd_searx(bot, _msg(f"!searx {long_query}")))
|
||||
assert "too long" in bot.replied[0]
|
||||
|
||||
def test_title_truncation(self):
|
||||
"""Long titles are truncated to _MAX_TITLE_LEN."""
|
||||
long_title = "A" * 100
|
||||
data = {"results": [
|
||||
{"title": long_title, "url": "http://x.com", "content": "s"},
|
||||
]}
|
||||
bot = _FakeBot()
|
||||
|
||||
async def inner():
|
||||
with patch("urllib.request.urlopen", return_value=_FakeResp(data)):
|
||||
await cmd_searx(bot, _msg("!searx test"))
|
||||
assert len(bot.replied) == 1
|
||||
title_part = bot.replied[0].split(" -- ")[0]
|
||||
assert len(title_part) <= 80
|
||||
assert title_part.endswith("...")
|
||||
|
||||
asyncio.run(inner())
|
||||
|
||||
def test_no_title_fallback(self):
|
||||
"""Empty title shows '(no title)'."""
|
||||
data = {"results": [
|
||||
{"title": "", "url": "http://x.com", "content": "s"},
|
||||
]}
|
||||
bot = _FakeBot()
|
||||
|
||||
async def inner():
|
||||
with patch("urllib.request.urlopen", return_value=_FakeResp(data)):
|
||||
await cmd_searx(bot, _msg("!searx test"))
|
||||
assert "(no title)" in bot.replied[0]
|
||||
|
||||
asyncio.run(inner())
|
||||
Reference in New Issue
Block a user