LMH-Detection-Framework: A Web-Based Pipeline for Detecting Linear Magnetic Holes in Wind MFI H4-RTN Data

A self-contained web application for the automated detection of Linear Magnetic Holes (LMHs) in magnetic-field data from the Wind spacecraft's Magnetic Field Investigation (MFI), using H4-RTN resolution CSV files. The framework provides a browser-based interface in which every detection parameter is exposed as a configurable form field, with a live streaming console log and one-click download of results. The underlying pipeline processes data in overlapping chunks, allowing multi-gigabyte input files to be analysed without loading them fully into memory. Detection proceeds by identifying local minima in the total magnetic field magnitude within a rolling window, refining event boundaries against an ambient baseline computed from regions before and after each candidate, and filtering candidates on two criteria: a depth ratio (B_min / B_ave) and the angular deviation of the ambient field vector across the event. Configurable parameters include chunk size, overlap buffer, rolling-minimum window, boundary search window, ambient region lengths, depth-ratio threshold, and angle threshold. For each input file the tool produces a CSV of detected events (start time, end time, time of minimum, duration, depth ratio, angular deviation, ambient field magnitude, and minimum field magnitude) together with a PDF report containing summary statistics and a per-event plot of the total field with the event region highlighted. Input data must contain ISO-8601 epoch timestamps and the BR, BT, and BN components in RTN coordinates. Wind MFI data are not distributed with this software and must be obtained separately from NASA's Coordinated Data Analysis Web (CDAWeb). Implemented in Python (Flask, NumPy, pandas, Matplotlib) with a self-bootstrapping launcher for Windows. Includes a sample output report (detected_lmh_events_2018.pdf) demonstrating detection results on 2018 Wind MFI data.