chore: remove Grid Operations Primer from report site

CLAUDE.md

@@ -0,0 +1,89 @@
+# CLAUDE.md — grc-tech-evaluation
+
+## Repo Overview
+
+Standalone evaluation workspace for comparing six power-system modeling tools.
+This is NOT an installable Python package — each `evaluations/<tool>/` directory
+is an independent project with its own virtualenv and dependency set.
+
+No ZGE internal dependencies. No connection to the trading platform repos.
+
+## Directory Layout
+
+- `evaluation_guides/` — Rubric and test protocol (read-only reference)
+- `data/networks/` — Shared MATPOWER .m case files used by all tools
+- `evaluations/<tool>/` — One directory per tool under evaluation
+- `evaluations/<tool>/results/<dimension>/` — Test outputs organized by rubric dimension
+
+## Execution Environment
+
+**Never run code, tests, linters, or pre-commit locally.** Always use the devcontainer.
+All commands (pytest, julia, octave, pre-commit, etc.) must run inside the container.
+
+Use the `dc-exec` helper (`.devcontainer/dc-exec`) which works from both the main
+checkout and any git worktree:
+
+```bash
+# Open a shell inside the devcontainer
+.devcontainer/dc-exec bash
+
+# Run a one-off command
+.devcontainer/dc-exec <command>
+
+# Run in a specific container directory
+.devcontainer/dc-exec -C /workspace/evaluations/pypsa uv run python -c "import pypsa"
+```
+
+`dc-exec` finds the container via `git worktree list` + Docker labels, so it works
+from `.claude/worktrees/<name>/` where `devcontainer exec --workspace-folder .` cannot.
+
+## Per-Tool Setup
+
+### Python tools (pypsa, pandapower, gridcal)
+
+```bash
+cd evaluations/<tool>
+uv sync
+uv run python verify_install.py
+```
+
+Do NOT use `pip install` — these are uv-managed projects.
+
+### Julia tools (powermodels, powersimulations)
+
+```bash
+cd evaluations/<tool>
+julia --project=. -e 'using Pkg; Pkg.instantiate()'
+julia --project=. verify_install.jl
+```
+
+**Julia startup is slow by design.** `Pkg.instantiate()` compiles packages to native code on first run
+(can take many minutes). Subsequent runs reuse the precompiled cache but still pay a 5–15s load
+tax per invocation to deserialize it.
+
+For repeated evaluation runs, stay in the REPL and `include()` your script instead of re-launching:
+
+```julia
+# start once
+julia --project=.
+
+# then inside the REPL, re-run without restart overhead:
+julia> include("my_eval_script.jl")
+```
+
+### MATPOWER (Octave)
+
+```bash
+cd evaluations/matpower
+bash setup.sh        # downloads MATPOWER 8.1
+octave verify_install.m
+```
+
+## Conventions
+
+- Python 3.12, Julia 1.10
+- Ruff for Python linting (line-length = 100)
+- Conventional commits enforced by pre-commit
+- Each tool has isolated dependencies — no shared virtualenv
+- Results go in `evaluations/<tool>/results/<dimension>/`
+- The seven rubric dimensions: gate, expressiveness, extensibility, scalability, accessibility, maturity, supply_chain

data/fnm/.gitignore

@@ -0,0 +1,28 @@
+# FNM Data Files — NDA-restricted, never commit
+*.raw
+*.RAW
+*.csv
+*.CSV
+*.parquet
+*.m
+
+# Intermediate and reference output directories
+intermediate/**
+!intermediate/
+!intermediate/README.md
+!intermediate/schemas/
+!intermediate/schemas/*.json
+reference/**
+!reference/
+!reference/*/
+!reference/pass_conditions.json
+
+# Allow tracked infrastructure files
+!manifest.json
+!README.md
+!**/README.md
+!scripts/**/*.py
+!scripts/**/*.m
+!docs/**/*.md
+!docs/**/*.json
+!.gitignore

data/fnm/README.md

@@ -0,0 +1,51 @@
+# FNM Data Directory
+
+This directory contains artifacts related to the Full Network Model (FNM) ingestion
+pipeline. The FNM data itself is NDA-restricted and **must never be committed to version
+control**.
+
+## Directory Layout
+
+```
+data/fnm/
+├── manifest.json          # Machine-readable list of expected FNM source files
+├── .gitignore             # Blocks all FNM data files from version control
+├── README.md              # This file
+├── intermediate/          # Parser output (Parquet, intermediate formats)
+├── reference/             # Phase 3 verification / reference solution datasets
+├── docs/                  # Reference documentation for PSS/E record types, supplemental CSVs
+└── scripts/               # Python modules for parsing, validation, and manifest I/O
+```
+
+## NDA Restrictions
+
+FNM source files (PSS/E RAW files and supplemental CSVs) are provided under NDA.
+They must not be committed, shared publicly, or stored in any unencrypted location outside
+approved infrastructure.
+
+The `.gitignore` in this directory blocks all data file extensions (`*.raw`, `*.csv`,
+`*.parquet`, `*.m`) and the `intermediate/` and `reference/` directories from being tracked.
+
+## FNM_PATH Environment Variable
+
+All FNM-dependent code is gated by the `FNM_PATH` environment variable. Set it to the
+directory containing the FNM source files:
+
+```bash
+export FNM_PATH=/path/to/fnm/source/files
+```
+
+When `FNM_PATH` is not set, FNM-dependent tests are skipped and parsers will raise errors
+if invoked directly.
+
+## Obtaining FNM Source Files
+
+FNM source files must be obtained through the ISO's authorized distribution channels.
+Contact the data engineering team for access. The `manifest.json` file lists all expected
+source files with descriptions.
+
+## Manifest
+
+The `manifest.json` file enumerates every expected FNM source file. Use the
+`scripts/manifest_io.py` module to load, validate, and update the manifest
+programmatically rather than editing the JSON directly.

data/fnm/__init__.py

@@ -0,0 +1 @@
+from __future__ import annotations

data/fnm/docs/README.md

@@ -0,0 +1,11 @@
+# FNM Documentation Directory
+
+This directory contains reference documentation for FNM data formats, PSS/E record types,
+and supplemental CSV schemas. Documentation here supports developers working with the
+FNM ingestion pipeline.
+
+## Contents (populated by later phases)
+
+- PSS/E v31 RAW format record type documentation
+- Supplemental CSV field descriptions and schemas
+- Data dictionary mapping FNM fields to internal representations