Files
tai/taicpp/tests/misc_diff.py
T

274 lines
11 KiB
Python

#!/usr/bin/env python3
"""Differential test of the misc modules: runs each scenario through the real tai.py (tests/misc_py.py)
and through the C++ driver (build/misc/misc_driver) on identical state, then diffs the transcripts and
the resulting file trees.
Usage (from taicpp/): python3 tests/misc_diff.py [scenario ...] (build first: make -f tests/misc.mk all)
Exit code 0 only when every selected scenario matches byte for byte after path/timestamp normalising.
"""
import difflib
import http.server
import json
import threading
import hashlib
import os
import re
import shutil
import subprocess
import sys
import tempfile
HERE = os.path.dirname(os.path.abspath(__file__))
TAICPP = os.path.dirname(HERE)
DRIVER = os.path.join(TAICPP, "build", "misc", "misc_driver")
PYDRV = os.path.join(HERE, "misc_py.py")
DATA = os.path.join(HERE, "misc_data")
ORDER = ["json", "datetime", "chunk", "fts", "skills_data", "skills_files", "skills_ops", "install_hooks", "install_hooks_noshell", "install_ops",
"install_noengine", "vault_line", "sched", "rag", "telegram", "telegram_install", "release", "audit"]
EXTRA_BIN = {"install_hooks": ("zsh", "fish")} # scenario -> extra executables to create in the restricted PATH dir
TZ = "Europe/Amsterdam"
TS_RE = re.compile(r"\d{4}-\d{2}-\d{2}T\d{2}:\d{2}:\d{2}(?:\.\d+)?(?:[+-]\d{2}:\d{2})?")
STAMP_RE = re.compile(r"\d{8}T\d{6}Z")
PID_RE = re.compile(r"\.tmp-\d+")
TG_UPDATES = [
{"update_id": 10, "message": {"chat": {"id": 111}, "text": "/start"}},
{"update_id": 11, "message": {"chat": {"id": 111}, "text": "hello there"}},
{"update_id": 12, "message": {"chat": {"id": 222}, "voice": {"file_id": "f1"}}},
{"update_id": 13, "message": {"chat": {"id": 111}, "text": "explode now"}},
{"update_id": 14},
{"update_id": 15, "message": {"chat": {"id": 333}, "text": "/new"}},
{"update_id": 16, "message": {"chat": {"id": 111}, "text": "long"}},
{"message": {"chat": {"id": 5}, "text": "no update id"}},
]
class TgServer:
"""Scripted Bot API on localhost; records every request so both sides can be compared on the wire."""
def __init__(self):
self.log = []
self.polls = {}
outer = self
class H(http.server.BaseHTTPRequestHandler):
def log_message(self, *a):
pass
def reply(self, code, body, reason=None):
data = body if isinstance(body, bytes) else body.encode()
self.send_response(code, reason)
self.send_header("Content-Type", "application/json")
self.send_header("Content-Length", str(len(data)))
self.end_headers()
self.wfile.write(data)
def handle_any(self, method):
n = int(self.headers.get("Content-Length") or 0)
body = self.rfile.read(n).decode() if n else ""
outer.log.append("%s %s ct=%s body=%s" % (method, self.path, self.headers.get("Content-Type"), body))
path = self.path
if path.startswith("/file/bot"):
if path.endswith("voice/f1.oga"):
return self.reply(200, b"AUDIO")
if path.endswith("voice/fmodel.oga"):
return self.reply(200, b"OTHER")
return self.reply(404, "{}", "Not Found")
token, _, name = path[len("/bot"):].partition("/")
if token == "bad":
return self.reply(401, json.dumps({"ok": False, "description": "Unauthorized"}), "Unauthorized")
if token == "nojson" and name == "getMe":
return self.reply(200, "not json")
if name == "getMe":
return self.reply(200, json.dumps({"ok": True, "result": {} if token == "noname" else {"username": "tai_test_bot"}}))
if name == "getUpdates":
count = outer.polls[token] = outer.polls.get(token, 0) + 1
if count == 1:
return self.reply(200, json.dumps({"ok": True, "result": TG_UPDATES}))
return self.reply(502, "{}", "Bad Gateway")
if name == "getFile":
fid = json.loads(body).get("file_id")
if fid == "nofile":
return self.reply(200, json.dumps({"ok": True, "result": {}}))
if fid == "fbad":
return self.reply(500, "{}", "Internal Server Error")
if fid == "fmodel":
return self.reply(200, json.dumps({"ok": True, "result": {"file_path": "voice/fmodel.oga"}}))
if fid == "fmissingfile":
return self.reply(200, json.dumps({"ok": True, "result": {"file_path": "voice/gone.oga"}}))
return self.reply(200, json.dumps({"ok": True, "result": {"file_path": "voice/f1.oga"}}))
return self.reply(200, json.dumps({"ok": True, "result": {}}))
def do_POST(self):
self.handle_any("POST")
def do_GET(self):
self.handle_any("GET")
self.httpd = http.server.ThreadingHTTPServer(("127.0.0.1", 0), H)
self.port = self.httpd.server_address[1]
threading.Thread(target=self.httpd.serve_forever, daemon=True).start()
def stop(self):
self.httpd.shutdown()
self.httpd.server_close()
SERVER_SCENARIOS = {"telegram", "telegram_install"}
def normalize(text, root):
text = text.replace(root, "<ROOT>")
text = TS_RE.sub("<TS>", text)
text = STAMP_RE.sub("<STAMP>", text)
text = PID_RE.sub(".tmp-<PID>", text)
return text
def make_root(base, name, extra_bins):
root = os.path.join(base, name)
for sub in ("home/.tai", "proj", "bin"):
os.makedirs(os.path.join(root, sub), exist_ok=True)
os.symlink(sys.executable, os.path.join(root, "bin", "python3"))
for tool in extra_bins:
path = os.path.join(root, "bin", tool)
with open(path, "w") as handle:
handle.write("#!/bin/sh\nexit 0\n")
os.chmod(path, 0o755)
return root
def env_for(root):
return {
"HOME": os.path.join(root, "home"),
"TAI_HOME": os.path.join(root, "home", ".tai"),
"PATH": os.path.join(root, "bin"),
"TZ": TZ,
"LANG": "C.UTF-8",
"MISC_DATA_DIR": DATA,
"PYTHONIOENCODING": "utf-8",
"SHELL": "/bin/bash",
"MISC_TAI_PY": os.path.join(os.path.dirname(TAICPP), "tai.py"),
}
def tree(root):
found = {}
for dirpath, dirs, files in os.walk(root):
dirs.sort()
rel_dir = os.path.relpath(dirpath, root)
if rel_dir == "bin" or rel_dir.startswith("bin" + os.sep):
continue
if rel_dir.startswith(os.path.join("home", ".tai", "venv")):
continue # venv contents embed the invoking interpreter path; the report lines are compared instead
for d in dirs:
full = os.path.join(dirpath, d)
if os.path.islink(full):
found[os.path.relpath(full, root)] = "symlink -> " + normalize(os.readlink(full), root)
for f in files:
full = os.path.join(dirpath, f)
rel = os.path.relpath(full, root)
if os.path.islink(full):
found[rel] = "symlink -> " + normalize(os.readlink(full), root)
continue
with open(full, "rb") as handle:
raw = handle.read()
mode = oct(os.stat(full).st_mode & 0o777)
try:
body = normalize(raw.decode("utf-8"), root)
digest = hashlib.sha256(body.encode("utf-8")).hexdigest()[:16]
except UnicodeDecodeError:
digest = "bin-" + hashlib.sha256(raw).hexdigest()[:16]
if rel.endswith((".db", ".db-wal", ".db-shm")):
digest = "db"
found[normalize(rel, root)] = "%s %s" % (mode, digest)
return found
def run(cmd, root, extra_env=None):
env = env_for(root)
if extra_env:
env.update(extra_env)
done = subprocess.run(cmd, cwd=os.path.join(root, "proj"), env=env, stdin=subprocess.DEVNULL, stdout=subprocess.PIPE, stderr=subprocess.PIPE,
start_new_session=True, timeout=600)
return done.returncode, done.stdout.decode("utf-8", "replace"), done.stderr.decode("utf-8", "replace")
def check(name, base, verbose):
root_py = make_root(base, name + "-py", EXTRA_BIN.get(name, ()))
code, _out, err = run([sys.executable, "-I", PYDRV, name, root_py, "prepare"], root_py)
if code != 0:
return False, "python prepare failed (%d): %s" % (code, err[-2000:])
root_cpp = os.path.join(base, name + "-cpp")
shutil.copytree(root_py, root_cpp, symlinks=True)
# copytree keeps absolute symlink targets; point the bin symlink at the same interpreter (already absolute)
def one(cmd, root):
if name not in SERVER_SCENARIOS:
return run(cmd, root)
srv = TgServer()
try:
extra = {"MISC_TG_API": "http://127.0.0.1:%d/bot" % srv.port, "MISC_TG_FILE": "http://127.0.0.1:%d/file/bot" % srv.port}
code, o, e = run(cmd, root, extra)
finally:
srv.stop()
o = o + "---- wire log\n" + "\n".join(l.replace(str(srv.port), "PORT") for l in srv.log)
if name == "telegram":
o += "\n---- stderr\n" + e
return code, o, ""
code_p, out_p, err_p = one([sys.executable, "-I", PYDRV, name, root_py], root_py)
code_c, out_c, err_c = one([DRIVER, name, root_cpp], root_cpp)
out_p = normalize(out_p, root_py)
out_c = normalize(out_c, root_cpp)
problems = []
if code_p != 0:
problems.append("python exit %d: %s" % (code_p, err_p[-2000:]))
if code_c != 0:
problems.append("cpp exit %d: %s" % (code_c, err_c[-2000:]))
if out_p != out_c:
diff = list(difflib.unified_diff(out_p.splitlines(), out_c.splitlines(), "python", "cpp", lineterm="", n=1))
problems.append("transcript differs (%d diff lines):\n%s" % (len(diff), "\n".join(diff[:60])))
tp, tc = tree(root_py), tree(root_cpp)
if tp != tc:
lines = []
for key in sorted(set(tp) | set(tc)):
if tp.get(key) != tc.get(key):
lines.append(" %s: py=%s cpp=%s" % (key, tp.get(key), tc.get(key)))
problems.append("file tree differs:\n%s" % "\n".join(lines[:60]))
if verbose:
sys.stdout.write("---- python transcript (%d lines)\n%s\n" % (len(out_p.splitlines()), out_p))
return not problems, "\n".join(problems), len(out_p.splitlines()), len(tp)
def main():
args = [a for a in sys.argv[1:] if not a.startswith("-")]
verbose = "-v" in sys.argv
if not os.path.isdir(DATA):
subprocess.run([sys.executable, "-I", os.path.join(HERE, "gen_misc_data.py")], check=True)
names = args or ORDER
base = tempfile.mkdtemp(prefix="misc-diff-")
failed = 0
try:
for name in names:
result = check(name, base, verbose)
ok, why = result[0], result[1]
if ok:
print("PASS %-14s transcript=%d lines, tree=%d files" % (name, result[2], result[3]))
else:
failed += 1
print("FAIL %s\n%s" % (name, why))
finally:
if "--keep" not in sys.argv:
shutil.rmtree(base, ignore_errors=True)
else:
print("kept %s" % base)
print("%d scenario(s), %d failed" % (len(names), failed))
return 1 if failed else 0
if __name__ == "__main__":
sys.exit(main())