Skip to main content

Module solversummary

Module solversummary 

Source
Expand description

A short summary of solver output files (Fluent transcripts and report files, Nastran .f06 and decks, STAR-CCM+ macros and batch logs): the solver and its version, how the run ended, the last iteration or time step, and counts of FATAL / WARNING messages.

Only the head and the tail of a file are read, at most WINDOW (64 KiB) each, never the middle of a large file. A count therefore covers the two windows only and is labelled that way (Summary::line). See docs/src/research/solver-files.md.

The rules are hypotheses. They follow public documentation and typical output; no real Fluent, Nastran or STAR-CCM+ output was available. A summary is a hint taken from a user-written file, never a verified result: the UI shows it as a line of text and the MCP tools hand it out as untrusted data.

The text is cut and cleaned (printable characters, bounded length), so a hostile file cannot make a summary large or inject markup.

Structs§

Summary
The summary of one file version; stored as JSON in file_versions.solver_summary.

Enums§

Status
How a run ended, as far as the output tells.

Constants§

WINDOW
Bytes of the head and of the tail that are looked at.

Functions§

applies
Whether summarize has rules for a file: its format id, or none for an unclassified or plain-text .log / .txt (a STAR-CCM+ batch log has no format).
summarize
Summarises a solver file from its head and tail (each cut to WINDOW); None when the content does not look like the format. format_id is the cadformat id; "" for an unclassified text file.
windows
The two windows of a file: the first WINDOW bytes of head and the last WINDOW bytes of the file’s end, starting at a line. tail is the end of a larger file (empty when head holds the whole file); then the part of head behind the first window serves as the end, so the two windows never overlap.