Geocrawler

What it is for

Exploration data accumulates faster than anyone catalogs it. A project folder inherited from a partner, a drive returned by a contractor, or twenty years of a company’s own work all arrive as the same thing: a file tree nobody can characterise without opening it. How much of this is seismic? How many wells? Is this the same survey as that one, under a different name? Which of these files claim a coordinate system, and should we believe them?

Geocrawler answers those questions by reading the data rather than the folder names. It crawls the tree, categorises what it finds, opens the headers, and produces a report you can hand to someone else.

What it reads

• SEG-Y — binary and textual headers, geometry, survey metadata, and the byte layouts that vary between vendors more than the standard suggests they should.
• LAS well logs — curves, headers and well identification, with format and status reported separately so an unsupported LAS 3.0 file is distinguishable from a supported file that failed to read.
• Coordinate reference systems — detected from the data, with EPSG codes and project context where they can be established.

Duplicates

Duplicate detection works at file level and at folder level. The folder-level check is the one that earns its keep: it is common to find the same survey delivered twice into differently named directories, and comparing content rather than names is what surfaces it. That is usually the single largest reclaimable chunk of a repository.

What comes out

A PySide6 desktop application for reviewing results interactively, and exports for everyone who was not at the screen: Excel workbooks for stakeholder review and handoff, and Markdown reports for quick documentation. The catalog data underneath supports ongoing data management rather than being a one-off summary.

Why it is careful about what it claims

A header is an assertion, not a fact. Files routinely carry coordinate systems that were never correct, byte layouts that do not match the standard they name, and survey names that disagree with their contents. Geocrawler reports what a file claims, marks what it inferred, and distinguishes “this format is not supported” from “this file failed to read” — because those lead to completely different next actions, and conflating them wastes an afternoon.

What it does not do