Public experimental source retrieval and extraction

Source: Ivanovich, N.; Marsili, E.; Shen, X.; Messinese, E.; Marcos;
Rajala, P.; Lauro, F.M. (2025). Exploring the impact of flow dynamics on
corrosive biofilms under simulated deep-sea high-pressure conditions using
bio-electrochemostasis. Frontiers in Microbiology 16,1540664.
https://doi.org/10.3389/fmicb.2025.1540664

The article's Data availability statement links the public author repository:
https://github.com/NicoloIvanovich/Exploring-the-impact-of-flow-dynamics-in-deep-sea-corrosive-biofilms

Reproduce the factual extraction from this directory:
  python -m pip install -r requirements-source.txt
  python retrieve_source.py
  python extract_source.py

The extraction was verified with Python 3.12.14 and openpyxl 3.1.5. Source
extraction requires only openpyxl plus Python's standard library; subsequent
model analysis has separate numerical dependencies in the full package.

retrieve_source.py downloads the fixed public raw ZIP URL and verifies its
expected SHA256 and byte count before opening it. It reads only the exact
member Raw data/Weight loss.xlsx, verifies that workbook's SHA256, and writes
the unchanged bytes to source/Weightloss.xlsx. It does not use extractall,
download other repository files, execute source code, or write the ZIP to disk.
If the author archive changes, it stops rather than silently accepting a new
version. An existing workbook with the expected hash is reused; a different
existing workbook is never overwritten.

extract_source.py reads that workbook without modifying it. It produces:
  Ivanovich2025_weight_loss_32_measured_coupons.csv
  Ivanovich2025_condition_summary_8.csv
  Ivanovich2025_biological_replicate_means_16.csv
  Ivanovich2025_extraction_audit.json

Every extracted response retains its original sheet and cell identifier. There
are 32 measured coupons nested in 16 treatment-by-biological-replicate units;
technical triplicate weight readings are not additional experimental samples.
Reported workbook rates and published condition means are preserved, including
one flagged difference between a cached pre-incubation average and the mean of
its listed technical readings. No unmeasured holdup or bulk SRB concentration
is assigned to the specimens.

The associated article carries CC BY 4.0. The author GitHub repository does not
provide a separate LICENSE for its raw archive. This package supplies the
retrieval/extraction scripts and attributed numerical facts and derived
results; it does not bundle the author's raw ZIP or workbooks or assert a new
license for them.
