|
| 1 | +# emacs: -*- mode: python; py-indent-offset: 4; indent-tabs-mode: nil -*- |
| 2 | +"""Database helpers used by the CLI and plotting utilities.""" |
| 3 | + |
| 4 | +from __future__ import annotations |
| 5 | + |
| 6 | +import datetime |
| 7 | +from typing import Tuple |
| 8 | + |
| 9 | +import pandas as pd |
| 10 | +from pymongo import MongoClient |
| 11 | + |
| 12 | + |
| 13 | +def load_event(event_name: str, unique: bool = True) -> pd.DataFrame: |
| 14 | + """Load one event collection from MongoDB.""" |
| 15 | + db = MongoClient().fmriprep_stats |
| 16 | + data = pd.DataFrame(list(db[event_name].find())) |
| 17 | + if len(data) == 0: |
| 18 | + raise RuntimeError(f"No records of event '{event_name}'") |
| 19 | + |
| 20 | + data["dateCreated"] = pd.to_datetime(data["dateCreated"]) |
| 21 | + data["date_minus_time"] = data["dateCreated"].apply( |
| 22 | + lambda df: datetime.datetime(year=df.year, month=df.month, day=df.day) |
| 23 | + ) |
| 24 | + if unique: |
| 25 | + data = data.drop_duplicates(subset=["run_uuid"]) |
| 26 | + return data |
| 27 | + |
| 28 | + |
| 29 | +def massage_versions( |
| 30 | + started: pd.DataFrame, success: pd.DataFrame |
| 31 | +) -> Tuple[pd.DataFrame, pd.DataFrame]: |
| 32 | + """Normalize version strings as done in the analysis notebook.""" |
| 33 | + started = started.copy() |
| 34 | + success = success.copy() |
| 35 | + |
| 36 | + started = started.fillna(value={"environment_version": "older"}) |
| 37 | + success = success.fillna(value={"environment_version": "older"}) |
| 38 | + |
| 39 | + started.loc[started.environment_version == "v0.0.1", "environment_version"] = "older" |
| 40 | + success.loc[success.environment_version == "v0.0.1", "environment_version"] = "older" |
| 41 | + |
| 42 | + started.loc[started.environment_version.str.startswith("20.0"), "environment_version"] = "older" |
| 43 | + success.loc[success.environment_version.str.startswith("20.0"), "environment_version"] = "older" |
| 44 | + started.loc[started.environment_version.str.startswith("20.1"), "environment_version"] = "older" |
| 45 | + success.loc[success.environment_version.str.startswith("20.1"), "environment_version"] = "older" |
| 46 | + |
| 47 | + versions = sorted( |
| 48 | + { |
| 49 | + ".".join(v.split(".")[:2]) |
| 50 | + for v in started.environment_version.unique() |
| 51 | + if "." in str(v) |
| 52 | + } |
| 53 | + ) |
| 54 | + for ver in versions: |
| 55 | + started.loc[started.environment_version.str.startswith(ver), "environment_version"] = ver |
| 56 | + success.loc[success.environment_version.str.startswith(ver), "environment_version"] = ver |
| 57 | + |
| 58 | + return started, success |
0 commit comments