mirror of
https://github.com/NicolasBohn/NexQuant.git
synced 2026-07-28 07:57:44 +00:00
you-n-g
4f484e009a
* Successfully logging the trace * Start debugging & Add policy file * Support loading sessions * Add docs * Add tqdm
143 lines
4.9 KiB
Python
143 lines
4.9 KiB
Python
import re
|
|
import json
|
|
import pickle
|
|
from datetime import datetime, timezone
|
|
from pathlib import Path
|
|
from typing import Literal, Generator
|
|
|
|
from .base import Message, Storage
|
|
|
|
|
|
class FileStorage(Storage):
|
|
"""
|
|
The info are logginged to the file systems
|
|
|
|
TODO: describe the storage format
|
|
"""
|
|
|
|
def __init__(self, path: str = "./log/") -> None:
|
|
self.path = Path(path)
|
|
self.path.mkdir(parents=True, exist_ok=True)
|
|
|
|
def log(
|
|
self,
|
|
obj: object,
|
|
name: str = "",
|
|
save_type: Literal["json", "text", "pkl"] = "text",
|
|
timestamp: datetime | None = None,
|
|
) -> Path:
|
|
# TODO: We can remove the timestamp after we implement PipeLog
|
|
if timestamp is None:
|
|
timestamp = datetime.now(timezone.utc)
|
|
else:
|
|
timestamp = timestamp.astimezone(timezone.utc)
|
|
|
|
cur_p = self.path / name.replace(".", "/")
|
|
cur_p.mkdir(parents=True, exist_ok=True)
|
|
|
|
path = cur_p / f"{timestamp.strftime('%Y-%m-%d_%H-%M-%S-%f')}.log"
|
|
|
|
if save_type == "json":
|
|
path = path.with_suffix(".json")
|
|
with path.open("w") as f:
|
|
try:
|
|
json.dump(obj, f)
|
|
except TypeError:
|
|
json.dump(json.loads(str(obj)), f)
|
|
return path
|
|
elif save_type == "pkl":
|
|
path = path.with_suffix(".pkl")
|
|
with path.open("wb") as f:
|
|
pickle.dump(obj, f)
|
|
return path
|
|
elif save_type == "text":
|
|
obj = str(obj)
|
|
with path.open("w") as f:
|
|
f.write(obj)
|
|
return path
|
|
|
|
log_pattern = re.compile(
|
|
r"(?P<timestamp>\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}\.\d{3}) \| "
|
|
r"(?P<level>DEBUG|INFO|WARNING|ERROR|CRITICAL) *\| "
|
|
r"(?P<caller>.+:.+:\d+) - "
|
|
)
|
|
|
|
def iter_msg(self, watch: bool = False) -> Generator[Message, None, None]:
|
|
msg_l = []
|
|
for file in self.path.glob("**/*.log"):
|
|
tag = '.'.join(str(file.relative_to(self.path)).replace("/", ".").split(".")[:-3])
|
|
pid = file.parent.name
|
|
|
|
with file.open("r") as f:
|
|
content = f.read()
|
|
|
|
matches, next_matches = self.log_pattern.finditer(content), self.log_pattern.finditer(content)
|
|
next_match = next(next_matches, None)
|
|
# NOTE: the content will be the text between `match` and `next_match`
|
|
for match in matches:
|
|
next_match = next(next_matches, None)
|
|
|
|
timestamp_str = match.group("timestamp")
|
|
timestamp = datetime.strptime(timestamp_str, "%Y-%m-%d %H:%M:%S.%f").replace(tzinfo=timezone.utc)
|
|
level = match.group("level")
|
|
caller = match.group("caller")
|
|
|
|
# Extract the message content
|
|
message_start = match.end()
|
|
message_end = next_match.start() if next_match else len(content)
|
|
message_content = content[message_start:message_end].strip()
|
|
|
|
m = Message(
|
|
tag=tag,
|
|
level=level,
|
|
timestamp=timestamp,
|
|
caller=caller,
|
|
pid_trace=pid,
|
|
content=message_content
|
|
)
|
|
|
|
if "Logging object in" in m.content:
|
|
absolute_p = m.content.split("Logging object in ")[1]
|
|
relative_p = "." + absolute_p.split(self.path.name)[1]
|
|
pkl_path = self.path / relative_p
|
|
with pkl_path.open("rb") as f:
|
|
m.content = pickle.load(f)
|
|
|
|
msg_l.append(m)
|
|
|
|
msg_l.sort(key=lambda x: x.timestamp)
|
|
for m in msg_l:
|
|
yield m
|
|
|
|
def truncate(self, time: datetime) -> None:
|
|
# any message later than `time` will be removed
|
|
for file in self.path.glob("**/*.log"):
|
|
|
|
with file.open("r") as f:
|
|
content = f.read()
|
|
|
|
new_content = ""
|
|
|
|
matches, next_matches = self.log_pattern.finditer(content), self.log_pattern.finditer(content)
|
|
|
|
next_match = next(next_matches, None)
|
|
for match in matches:
|
|
next_match = next(next_matches, None)
|
|
timestamp_str = match.group("timestamp")
|
|
timestamp = datetime.strptime(timestamp_str, "%Y-%m-%d %H:%M:%S.%f").replace(tzinfo=timezone.utc)
|
|
|
|
log_start = match.start()
|
|
log_end = next_match.start() if next_match else len(content)
|
|
msg = content[match.end():log_end].strip()
|
|
|
|
if timestamp > time:
|
|
if "Logging object in" in msg:
|
|
absolute_p = msg.split("Logging object in ")[1]
|
|
p = Path(absolute_p)
|
|
p.unlink()
|
|
continue
|
|
|
|
new_content += content[log_start:log_end]
|
|
with file.open("w") as f:
|
|
f.write(new_content)
|