unitoolkit 0.7.0__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- ladok_vakt/__init__.py +7 -0
- ladok_vakt/__main__.py +51 -0
- ladok_vakt/core.py +553 -0
- ladok_vakt/login.py +232 -0
- tools.json +623 -0
- unitoolkit/__init__.py +1 -0
- unitoolkit/__main__.py +63 -0
- unitoolkit/agent.py +270 -0
- unitoolkit/browser.py +147 -0
- unitoolkit/canvas.py +189 -0
- unitoolkit/chainproxy.py +113 -0
- unitoolkit/citation.py +165 -0
- unitoolkit/databases.py +497 -0
- unitoolkit/docs.py +394 -0
- unitoolkit/fulltext.py +260 -0
- unitoolkit/koha.py +729 -0
- unitoolkit/mcp.py +85 -0
- unitoolkit/programme.py +147 -0
- unitoolkit/search.py +507 -0
- unitoolkit/serve.py +153 -0
- unitoolkit/sharepoint.py +86 -0
- unitoolkit/surface.py +86 -0
- unitoolkit/timeedit.py +209 -0
- unitoolkit/tools.py +555 -0
- unitoolkit-0.7.0.dist-info/METADATA +177 -0
- unitoolkit-0.7.0.dist-info/RECORD +30 -0
- unitoolkit-0.7.0.dist-info/WHEEL +5 -0
- unitoolkit-0.7.0.dist-info/entry_points.txt +4 -0
- unitoolkit-0.7.0.dist-info/licenses/LICENSE +21 -0
- unitoolkit-0.7.0.dist-info/top_level.txt +2 -0
unitoolkit/mcp.py
ADDED
|
@@ -0,0 +1,85 @@
|
|
|
1
|
+
"""Minimal MCP server over stdio (JSON-RPC 2.0, newline-delimited). Stdlib only.
|
|
2
|
+
|
|
3
|
+
python -m unitoolkit.mcp
|
|
4
|
+
|
|
5
|
+
Register it in any MCP client, e.g. Claude Code:
|
|
6
|
+
claude mcp add unitoolkit -- python -m unitoolkit.mcp
|
|
7
|
+
Hermes (config.yaml):
|
|
8
|
+
mcp_servers: {unitoolkit: {command: python3, args: [-m, unitoolkit.mcp]}}
|
|
9
|
+
|
|
10
|
+
Lists the question-shaped core tools by default (agent.CORE); UNITOOLKIT_TOOLS=all lists all,
|
|
11
|
+
or give a comma list. Every tool is callable by name regardless of what is listed.
|
|
12
|
+
"""
|
|
13
|
+
import json
|
|
14
|
+
import sys
|
|
15
|
+
import traceback
|
|
16
|
+
|
|
17
|
+
from . import __version__
|
|
18
|
+
from .tools import TOOLS, call, manifest
|
|
19
|
+
|
|
20
|
+
PROTOCOL = "2025-06-18"
|
|
21
|
+
|
|
22
|
+
|
|
23
|
+
def _reply(mid, result=None, error=None):
|
|
24
|
+
msg = {"jsonrpc": "2.0", "id": mid}
|
|
25
|
+
if error is not None:
|
|
26
|
+
msg["error"] = error
|
|
27
|
+
else:
|
|
28
|
+
msg["result"] = result
|
|
29
|
+
sys.stdout.write(json.dumps(msg, ensure_ascii=False) + "\n")
|
|
30
|
+
sys.stdout.flush()
|
|
31
|
+
|
|
32
|
+
|
|
33
|
+
def handle(req):
|
|
34
|
+
method, mid, params = req.get("method"), req.get("id"), req.get("params") or {}
|
|
35
|
+
if method == "initialize":
|
|
36
|
+
return {"protocolVersion": params.get("protocolVersion", PROTOCOL),
|
|
37
|
+
"capabilities": {"tools": {"listChanged": False}},
|
|
38
|
+
"serverInfo": {"name": "unitoolkit", "version": __version__},
|
|
39
|
+
"instructions": ("Read-only access to a student's own school systems: Canvas, "
|
|
40
|
+
"Ladok, the library, the timetable, and the subscribed research "
|
|
41
|
+
"databases. Start with `today` for 'what do I need to do', "
|
|
42
|
+
"`find_sources` for 'find me sources'. When anything fails, run "
|
|
43
|
+
"`doctor`: it names the fix, which is always a human action "
|
|
44
|
+
"(sign in once, renew a token). Never type a password. "
|
|
45
|
+
"Set UNITOOLKIT_TOOLS=all to list every tool.")}
|
|
46
|
+
if method == "tools/list":
|
|
47
|
+
from .agent import listed
|
|
48
|
+
return {"tools": manifest(listed(list(TOOLS)))}
|
|
49
|
+
if method == "tools/call":
|
|
50
|
+
name = params.get("name")
|
|
51
|
+
try:
|
|
52
|
+
out = call(name, params.get("arguments") or {})
|
|
53
|
+
return {"content": [{"type": "text",
|
|
54
|
+
"text": json.dumps(out, ensure_ascii=False, indent=1, default=str)}],
|
|
55
|
+
"isError": False}
|
|
56
|
+
except Exception as e: # tool errors are results, not protocol errors
|
|
57
|
+
return {"content": [{"type": "text", "text": f"{type(e).__name__}: {e}"}], "isError": True}
|
|
58
|
+
if method == "ping":
|
|
59
|
+
return {}
|
|
60
|
+
raise LookupError(method)
|
|
61
|
+
|
|
62
|
+
|
|
63
|
+
def main():
|
|
64
|
+
for line in sys.stdin:
|
|
65
|
+
line = line.strip()
|
|
66
|
+
if not line:
|
|
67
|
+
continue
|
|
68
|
+
try:
|
|
69
|
+
req = json.loads(line)
|
|
70
|
+
except ValueError:
|
|
71
|
+
_reply(None, error={"code": -32700, "message": "parse error"})
|
|
72
|
+
continue
|
|
73
|
+
if "id" not in req: # notification (e.g. notifications/initialized)
|
|
74
|
+
continue
|
|
75
|
+
try:
|
|
76
|
+
_reply(req["id"], handle(req))
|
|
77
|
+
except LookupError as e:
|
|
78
|
+
_reply(req["id"], error={"code": -32601, "message": f"method not found: {e}"})
|
|
79
|
+
except Exception as e:
|
|
80
|
+
_reply(req["id"], error={"code": -32603, "message": str(e),
|
|
81
|
+
"data": traceback.format_exc()[-800:]})
|
|
82
|
+
|
|
83
|
+
|
|
84
|
+
if __name__ == "__main__":
|
|
85
|
+
main()
|
unitoolkit/programme.py
ADDED
|
@@ -0,0 +1,147 @@
|
|
|
1
|
+
"""Build the programme map: every course, its kursplan and kurslitteraturlista, downloaded.
|
|
2
|
+
|
|
3
|
+
Source: the public course-overview page on shh.se. Course names are plain text, and the
|
|
4
|
+
document links appear as unordered "Kursplan"/"Kurslitteraturlista" anchor pairs, so they
|
|
5
|
+
are NOT paired by position. Instead every document is downloaded and named by the course
|
|
6
|
+
code found inside it (the page's own index misaligns whenever a course omits a link).
|
|
7
|
+
|
|
8
|
+
Documents live in SharePoint as share links; `download.aspx?share=<docid>` returns the
|
|
9
|
+
real bytes to a logged-in browser (`unitoolkit.browser`). SHH publishes kursplaner as HTML
|
|
10
|
+
pages and kurslitteraturlistor as PDFs.
|
|
11
|
+
|
|
12
|
+
python -m unitoolkit.programme
|
|
13
|
+
"""
|
|
14
|
+
import asyncio
|
|
15
|
+
import json
|
|
16
|
+
import pathlib
|
|
17
|
+
import re
|
|
18
|
+
import subprocess
|
|
19
|
+
|
|
20
|
+
from . import browser
|
|
21
|
+
from .sharepoint import fetch_share
|
|
22
|
+
|
|
23
|
+
LANDING = "https://www.shh.se/sv/utbildningar/grundniva/sjukskoterskeprogrammet-kurser/"
|
|
24
|
+
OUT = pathlib.Path("~/courses/programme").expanduser()
|
|
25
|
+
CODE_RE = re.compile(r"\b([0-9][A-ZÅÄÖ]{2,4}[0-9]+)\b")
|
|
26
|
+
COURSE_RE = re.compile(r"^(.+?),?\s*\(?([0-9][A-ZÅÄÖ]{2,4}[0-9]+)\)?,?\s*([\d.,]+)\s*hp\s*$")
|
|
27
|
+
|
|
28
|
+
|
|
29
|
+
async def _scrape(t):
|
|
30
|
+
from .browser import _eval, _session
|
|
31
|
+
|
|
32
|
+
async def steps(call):
|
|
33
|
+
await call("Page.enable")
|
|
34
|
+
await call("Page.navigate", url=LANDING)
|
|
35
|
+
await asyncio.sleep(4)
|
|
36
|
+
anchors = await _eval(call, (
|
|
37
|
+
"JSON.stringify([...document.querySelectorAll('a')]"
|
|
38
|
+
".filter(a=>/^Kurs(plan|litteraturlista)$/.test(a.innerText.trim()))"
|
|
39
|
+
".map(a=>[a.innerText.trim(),a.href]))"))
|
|
40
|
+
text = await _eval(call, "document.body.innerText")
|
|
41
|
+
return json.loads(anchors), text
|
|
42
|
+
|
|
43
|
+
return await _session(t["webSocketDebuggerUrl"], steps)
|
|
44
|
+
|
|
45
|
+
|
|
46
|
+
def _courses_from_text(text):
|
|
47
|
+
out = []
|
|
48
|
+
for line in text.split("\n"):
|
|
49
|
+
line = line.strip()
|
|
50
|
+
m = COURSE_RE.match(line)
|
|
51
|
+
if m:
|
|
52
|
+
out.append({"name": m.group(1).strip(), "code": m.group(2),
|
|
53
|
+
"hp": float(m.group(3).replace(",", "."))})
|
|
54
|
+
elif re.search(r"[0-9][A-ZÅÄÖ]{2,4}[0-9]", line):
|
|
55
|
+
code = CODE_RE.search(line)
|
|
56
|
+
out.append({"name": re.sub(r"\(?[0-9][A-ZÅÄÖ]{2,4}[0-9]+\)?,?.*$", "", line).strip(),
|
|
57
|
+
"code": code.group(1), "hp": None})
|
|
58
|
+
return out
|
|
59
|
+
|
|
60
|
+
|
|
61
|
+
def doc_text(path):
|
|
62
|
+
p = pathlib.Path(path)
|
|
63
|
+
if p.suffix in (".html", ".htm"):
|
|
64
|
+
raw = p.read_text(errors="ignore")
|
|
65
|
+
return re.sub(r"\s+", " ", re.sub(r"<[^>]+>", " ", raw))
|
|
66
|
+
if p.suffix == ".pdf":
|
|
67
|
+
try:
|
|
68
|
+
return subprocess.run(["pdftotext", "-q", str(p), "-"], capture_output=True,
|
|
69
|
+
text=True, timeout=60).stdout
|
|
70
|
+
except (OSError, subprocess.SubprocessError):
|
|
71
|
+
return ""
|
|
72
|
+
if p.suffix == ".docx":
|
|
73
|
+
try:
|
|
74
|
+
import html
|
|
75
|
+
import zipfile
|
|
76
|
+
x = zipfile.ZipFile(p).read("word/document.xml").decode("utf8", "ignore")
|
|
77
|
+
return html.unescape(re.sub(r"<[^>]+>", " ", x.replace("</w:p>", "\n")))
|
|
78
|
+
except Exception:
|
|
79
|
+
return ""
|
|
80
|
+
return ""
|
|
81
|
+
|
|
82
|
+
|
|
83
|
+
KNOWN_CODES = set()
|
|
84
|
+
|
|
85
|
+
|
|
86
|
+
def build(out=None):
|
|
87
|
+
global KNOWN_CODES, OUT
|
|
88
|
+
if out:
|
|
89
|
+
OUT = pathlib.Path(out).expanduser()
|
|
90
|
+
t = browser.new_tab(LANDING)
|
|
91
|
+
try:
|
|
92
|
+
anchors, text = asyncio.run(_scrape(t))
|
|
93
|
+
finally:
|
|
94
|
+
browser.close_tab(t["id"])
|
|
95
|
+
courses = _courses_from_text(text)
|
|
96
|
+
KNOWN_CODES = {c["code"] for c in courses}
|
|
97
|
+
OUT.mkdir(parents=True, exist_ok=True)
|
|
98
|
+
raw = OUT / "raw"
|
|
99
|
+
raw.mkdir(exist_ok=True)
|
|
100
|
+
got = []
|
|
101
|
+
for i, (kind, share) in enumerate(anchors):
|
|
102
|
+
body, ext = fetch_share(share)
|
|
103
|
+
if not body:
|
|
104
|
+
got.append({"kind": kind, "share": share, "error": "download failed"})
|
|
105
|
+
continue
|
|
106
|
+
kind_slug = "kursplan" if kind == "Kursplan" else "litteratur"
|
|
107
|
+
tmp = raw / f"_tmp_{i}.{ext}"
|
|
108
|
+
tmp.write_bytes(body)
|
|
109
|
+
text_ = doc_text(tmp)
|
|
110
|
+
codes = CODE_RE.findall(text_)
|
|
111
|
+
known = [c for c in codes if c in KNOWN_CODES]
|
|
112
|
+
code = (known or codes or [f"unknown{i}"])[0]
|
|
113
|
+
final = raw / f"{code}_{kind_slug}.{ext}"
|
|
114
|
+
if final.exists() and final.stat().st_size == tmp.stat().st_size:
|
|
115
|
+
tmp.unlink()
|
|
116
|
+
else:
|
|
117
|
+
tmp.rename(final)
|
|
118
|
+
got.append({"kind": kind, "code": code, "file": str(final), "bytes": len(body)})
|
|
119
|
+
for c in courses:
|
|
120
|
+
c["kursplan_file"] = next((g["file"] for g in got
|
|
121
|
+
if g.get("code") == c["code"] and g["kind"] == "Kursplan"), None)
|
|
122
|
+
c["litteratur_file"] = next((g["file"] for g in got
|
|
123
|
+
if g.get("code") == c["code"] and g["kind"] == "Kurslitteraturlista"), None)
|
|
124
|
+
(OUT / "programme.json").write_text(json.dumps({"courses": courses, "documents": got},
|
|
125
|
+
ensure_ascii=False, indent=1))
|
|
126
|
+
return courses, got
|
|
127
|
+
|
|
128
|
+
|
|
129
|
+
def report(courses):
|
|
130
|
+
lines = ["# Sjuksköterskeprogrammet 180 hp, kurskarta", "",
|
|
131
|
+
f"{len(courses)} kurser enligt shh.se. Kursplan och kurslitteraturlista hämtade från "
|
|
132
|
+
"programmets kursöversikt.", ""]
|
|
133
|
+
for i, c in enumerate(courses, 1):
|
|
134
|
+
lines.append(f"{i}. **{c['name']}** ({c['code']}): "
|
|
135
|
+
f"kursplan {'ja' if c.get('kursplan_file') else 'NEJ'}, "
|
|
136
|
+
f"litteratur {'ja' if c.get('litteratur_file') else 'NEJ'}")
|
|
137
|
+
return "\n".join(lines)
|
|
138
|
+
|
|
139
|
+
|
|
140
|
+
if __name__ == "__main__":
|
|
141
|
+
cs, docs = build()
|
|
142
|
+
md = report(cs)
|
|
143
|
+
(OUT / "KURSKARTA.md").write_text(md)
|
|
144
|
+
print(md)
|
|
145
|
+
print(f"\n{len(cs)} kurser, {sum(1 for c in cs if c.get('kursplan_file'))} kursplaner, "
|
|
146
|
+
f"{sum(1 for c in cs if c.get('litteratur_file'))} litteraturlistor, "
|
|
147
|
+
f"{sum(1 for d in docs if d.get('error'))} fel -> {OUT}")
|