Skip to content

Database Management

Overview

The Database Management tab builds custom NPZ libraries from a CSV of SMILES strings for use in Molecular Analysis → Load Other Database. Main capabilities:

  • Display built-in database sizes (Plant, Human, Microbial, All, etc.).
  • Background database generation via VirDBGenerator (NMR shift prediction and packaging).
  • Model file check for nmr_model.pt before allowing creation.
  • Progress bar and status log during long runs.

Built-in libraries (Database/*.npz) are shipped with the application and are not recreated on this tab.

Interface Layout

Single-column layout (no right results split):

Section Content
Database Statistics Bullet list of NP counts per built-in library
Create New Database SMILES CSV path, output .npz path, Create Database button
Progress & Status Model status label, progress bar, read-only log (QTextEdit)

Workflow

① Verify model status

  • On tab open, Model Status shows:
  • ✓ Available — VirMolAnalyte/VirDBcreator/NMRprediction/model/nmr_model.pt found
  • ✗ Missing — Create Database disabled

Packaged installs must include this file under the same relative path.

② Prepare input CSV

  • CSV with at least one column of SMILES strings (one compound per row).
  • Invalid SMILES are skipped by the generator (see log).

③ Set input and output paths

Field Description
SMILES File Browse to input .csv
Output Path Target .npz; if empty, default is GUI_result_files/generated_virDB.npz

Use the browse icons to select paths.

④ Generate database

  • Click Create Database.
  • DatabaseCreationThread runs VirDBGenerator from the project root working directory.
  • Watch Progress and Status Log (may take a long time for large CSVs).
  • On success, log shows absolute output path and notes the file can be loaded for identification.

⑤ Use the new library

  1. Open Molecular Analysis.
  2. Browse Other Database → select your .npz.
  3. Load Other Database → Start Analysis.

⑥ Output file structure

GUI_result_files/
  generated_virDB.npz    # Default output if path not specified

# Or user-selected path, e.g.
Database/my_lab.npz

Internal NPZ structure follows VirDBGenerator / DataPrepare conventions (same family as built-in Plant.npz).

FAQ

  1. Create Database disabled

Install or restore nmr_model.pt under VirMolAnalyte/VirDBcreator/NMRprediction/model/.

  1. Generation failed in log

Check SMILES validity, disk space, and write permissions on the output folder.

  1. Load Other Database fails on Molecular Analysis

NPZ must be produced by this tool (or compatible VirDBGenerator output), not an arbitrary archive.

  1. Built-in vs custom library

Built-in files live in Database/; custom files can live anywhere if you browse to them.

  1. Working directory messages in log

The thread temporarily chdirs to the app root; this is normal and restored after the job.

Relation to Other Modules

  • Molecular Analysis: Load Other Database consumes NPZ files created here.
  • Fragment Analysis: Uses screening hits from any loaded library; no direct tab link.
  • AI Assistant: Indirect — better hits improve LLM context.
  • Other Tool: Independent.