Documentation Index Fetch the complete documentation index at: https://mintlify.com/acdeveloper-sci/forti4d/llms.txt
Use this file to discover all available pages before exploring further.
Every time forti4d runs, it writes all output files to the directory configured via --output (or the FORT_OUT environment variable), which defaults to results/. Each pipeline step writes its own files and never modifies another step’s output, so you can safely re-run individual steps without disturbing the rest. All output files are written even when the corpus produces no data for them — in that case the file contains column headers only.
Output Directory Structure <FORT_OUT>/ ├── inventory_report.csv ├── dep_00_ambiguities.csv ├── dep_01_master_data.csv ├── dep_02_unit_graph.csv ├── dep_03_impact_matrix.csv ├── dep_04_external_orphans.csv ├── dep_05_file_dependencies.csv ├── dep_06_include_files.csv ├── report_density.csv ├── audit/ │ └── <file>_DEBUG.csv (one per source file) ├── blocks/ │ └── <file>_blocks.txt (one per source file) ├── report_structure_analysis.csv ├── report_migration_strategy.csv ├── PROJECT_SUMMARY.md ├── file_statistics.csv ├── report_complexity.csv ├── common_usage.csv ├── common_coupling.csv ├── symbol_variables.csv ├── symbol_signatures.csv ├── symbol_implicit.csv ├── type_definitions.csv ├── type_components.csv ├── equivalences.csv ├── report_reachability.csv ├── report_sloc.csv ├── report_clones.csv ├── report_consolidated.csv ├── graph_*.dot ├── report_prioritization.csv ├── report.html └── forti4d.log
audit/ and blocks/ are subdirectories, not flat files. Each contains one file per Fortran source file processed.
All Output Files Foundation
Profiling & Structure
Extended Analysis
E4 ScopeManager
Graphs & Reports
These files are produced by the first two pipeline steps and form the base that every subsequent script reads. File Produced by Description inventory_report.csvinventory.pyAll program units with type, parent unit, line range, and audit flags (legacy constructs, I/O statements). The foundation for all other scripts. dep_00_ambiguities.csvdependencies.pyUnits defined in more than one source file. dep_01_master_data.csvdependencies.pyRaw CALL, USE, and function-call relationships before resolution. dep_02_unit_graph.csvdependencies.pyResolved call graph with typed edges (CALL, USE, FUNC_CALL). dep_03_impact_matrix.csvdependencies.pyFan-In and Fan-Out per unit. dep_04_external_orphans.csvdependencies.pyExternal references with no known definition in the corpus. dep_05_file_dependencies.csvdependencies.pyFile-level dependency summary. dep_06_include_files.csvdependencies.pyFortran INCLUDE file references with existence status.
These files describe the internal character of each unit — what kinds of statements it contains and how it fits into the overall architecture. File Produced by Description report_density.csvprofiler.pyStatement density profile per unit: % calculation, % control flow, % I/O, % legacy, % declarations. audit/<file>_DEBUG.csvprofiler.pyPer-line statement classification for each source file. The primary intermediate artifact consumed by all extended analysis scripts. blocks/<file>_blocks.txtblock_analysis.pyBlock topology per source file — hierarchical view of control-flow nesting (IF/DO/SELECT depth) per unit. report_structure_analysis.csvstructure_analysis.pyArchitectural role per source file: CRITICAL_NODE, ORCHESTRATOR, ENTRY_POINT, WORKER, ISLAND, or MIXED. report_migration_strategy.csvcross_analysis.pyMigration strategy per unit — one of 7 recommended actions based on ICM and IVC indices. PROJECT_SUMMARY.mdexecutive_summary.pyExecutive summary in Markdown: global LOC, unit counts, type distribution, top monolithic files, and scope health indicators. file_statistics.csvexecutive_summary.pyPer-file summary statistics.
These files add deeper metrics on top of the foundation tier. They all require audit/*_DEBUG.csv files produced by profiler.py. File Produced by Description report_complexity.csvcomplexity.pyMcCabe cyclomatic complexity (CC) per unit, with a level label: LOW (1–10), MEDIUM (11–20), HIGH (21–50), CRITICAL (>50). common_usage.csvcommon_blocks.pyF77 COMMON block usage — which units share each block. common_coupling.csvcommon_blocks.pyCOMMON block coupling risk: LOW (1 unit), MEDIUM (2–4 units), HIGH (5+ units). report_reachability.csvreachability.pyReachability from all entry points via BFS: ENTRY_POINT, REACHABLE, or UNREACHABLE. report_sloc.csvsloc.pyPrecise SLOC count per unit: LOC, physical SLOC, net SLOC (logical statements only), blank lines, comments, continuations, comment density %. report_clones.csvclones.pyDuplicate unit pairs across files, classified as IDENTICAL, SIMILAR, or DIVERGED. report_consolidated.csvconsolidate.pyAll metrics joined into a single CSV — one row per unit, 34 columns. The primary analysis surface for manual investigation.
The E4 ScopeManager scripts extract the symbol-level microstructure of each unit (Axis Z of the MI4D model). File Produced by Description symbol_variables.csvsymbols.pyVariable and PARAMETER constant declarations per unit (type, name, attributes, COMMON membership). symbol_signatures.csvsymbols.pyFormal arguments of subroutines and functions with their declared types. symbol_implicit.csvsymbols.pyIMPLICIT rules per unit — records whether IMPLICIT NONE is in effect. type_definitions.csvderived_types.pyDerived TYPE definitions with host unit and component count. type_components.csvderived_types.pyComponent fields of each derived TYPE. equivalences.csvequivalences.pyEQUIVALENCE aliasing groups (union-find) — one row per variable per aliasing group.
Final-stage outputs consumed directly by engineers and project managers. File Produced by Description graph_complete.dotvisual_graph.pyFull corpus call graph — all nodes and all edges including USE. graph_simple.dotvisual_graph.pySimplified graph — reachable nodes only, CALL and FUNC_CALL edges only. graph_<names>.dotvisual_graph.pySubgraph scoped to one or more entry points (generated with --entry). report_prioritization.csvprioritization.pyUnits ranked by composite migration risk score (0–100), with priority tier and per-component scores. report.htmlhtml_report.pySelf-contained HTML report — filterable and sortable unit table, priority summary cards. No external dependencies. forti4d.logpipeline.pyFull-detail DEBUG log for the entire run. Written regardless of --quiet.
Reading report_consolidated.csv report_consolidated.csv is the central analysis surface — a single flat table joining every metric produced across the pipeline. It has one row per program unit and 34 columns spanning physical size, complexity, coupling, density, reachability, COMMON block usage, and the full E4 symbol summary.
Size columns LOC, SLOC_physical, SLOC_net, N_Comments, N_Continuation, Pct_Comment
Complexity & coupling CC, CC_Level, CC_SLOC, Fan_In, Fan_Out
Density profile Pct_Calc, Pct_Control, Pct_IO, Pct_Legacy
Reachability Status, Via_Entry_Points
COMMON blocks N_Common_Blocks, Common_Blocks
E4 scope N_Local_Vars, N_Formal_Args, Implicit_None, Has_Equiv, N_Equiv_Groups, N_Derived_Types
Open report_consolidated.csv in a spreadsheet. Sort by any column, add computed columns (e.g. CC × Fan_In), and filter to focus on specific risk profiles. The Interpretation guide explains exactly which combinations to look for.
Pipeline Order and Dependencies The diagram below shows which files feed which scripts. Output files are listed next to the arrow from each producer.
Fortran source files │ ▼ inventory.py → inventory_report.csv │ ├──► dependencies.py → dep_00…dep_06_*.csv │ ├──► profiler.py → report_density.csv │ └──► audit/*_DEBUG.csv │ │ │ ┌─────────┴──────────┐ │ ▼ ▼ │ block_analysis.py complexity.py │ common_blocks.py symbols.py │ equivalences.py derived_types.py │ ├──► sloc.py → report_sloc.csv ├──► clones.py → report_clones.csv ├──► structure_analysis.py → report_structure_analysis.csv ├──► cross_analysis.py → report_migration_strategy.csv ├──► reachability.py → report_reachability.csv │ [all reports above] │ ▼ consolidate.py → report_consolidated.csv │ ├──► visual_graph.py → graph_*.dot ├──► prioritization.py → report_prioritization.csv └──► html_report.py → report.html
consolidate.py must run after all analysis scripts. prioritization.py must run after consolidate.py. html_report.py must run after prioritization.py. The forti4d CLI manages this order automatically.
