refactor, security updates, cv extraction upgrades

2026-04-11 01:34:32 +02:00
parent 806b200ac5
commit 27fd70a2d7
59 changed files with 6817 additions and 1561 deletions
@@ -0,0 +1,91 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+ROOT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
+CORPUS_DIR="${CV_CORPUS_DIR:-$HOME/cvs}"
+TIMESTAMP="$(date +%Y%m%d-%H%M%S)"
+EXPORT_ROOT="${CV_JSON_EXPORT_DIR:-$ROOT_DIR/tmp/cv-json-export/$TIMESTAMP}"
+OUTPUT_ROOT="$EXPORT_ROOT/benchmark"
+APPROVED_DIR="$EXPORT_ROOT/approved-fixtures"
+EDITABLE_DIR="$EXPORT_ROOT/editable-json"
+DEFAULT_IGNORE_PATTERNS="cv-template.pdf,Resume.en.pdf,EPS-*.pdf"
+IGNORE_PATTERNS="${CV_BENCHMARK_IGNORE:-$DEFAULT_IGNORE_PATTERNS}"
+
+resolve_dotnet() {
+  if command -v dotnet >/dev/null 2>&1; then
+    command -v dotnet
+    return 0
+  fi
+
+  local candidates=(
+    "$HOME/.gsd/agent/bin/dotnet"
+    "$HOME/.dotnet/dotnet"
+    "/usr/bin/dotnet"
+    "/usr/local/bin/dotnet"
+  )
+
+  local candidate
+  for candidate in "${candidates[@]}"; do
+    if [[ -x "$candidate" ]]; then
+      printf '%s\n' "$candidate"
+      return 0
+    fi
+  done
+
+  return 1
+}
+
+DOTNET_BIN="$(resolve_dotnet || true)"
+
+if [[ ! -d "$CORPUS_DIR" ]]; then
+  echo "CV corpus directory not found: $CORPUS_DIR" >&2
+  exit 1
+fi
+
+if [[ -z "$DOTNET_BIN" ]]; then
+  echo "dotnet not found on PATH or known fallback locations." >&2
+  echo "Checked: ~/.gsd/agent/bin/dotnet, ~/.dotnet/dotnet, /usr/bin/dotnet, /usr/local/bin/dotnet" >&2
+  exit 1
+fi
+
+if [[ "$CORPUS_DIR" != "/home/pi/cvs" ]]; then
+  echo "This wrapper currently relies on the existing benchmark harness, which scans /home/pi/cvs." >&2
+  echo "Set up a symlink or move the corpus there, or rerun with CV_CORPUS_DIR=/home/pi/cvs." >&2
+  exit 1
+fi
+
+mkdir -p "$OUTPUT_ROOT" "$APPROVED_DIR" "$EDITABLE_DIR"
+
+echo "Exporting structured CV JSON from: $CORPUS_DIR"
+echo "Working directory: $EXPORT_ROOT"
+echo "Ignoring files: $IGNORE_PATTERNS"
+
+CV_BENCHMARK_OUTPUT_DIR="$OUTPUT_ROOT" \
+CV_BENCHMARK_APPROVED_DIR="$APPROVED_DIR" \
+CV_BENCHMARK_IGNORE="$IGNORE_PATTERNS" \
+  "$DOTNET_BIN" test "$ROOT_DIR/JobTrackerApi.Tests/JobTrackerApi.Tests.csproj" --filter CvCorpusHarnessTests /p:DisableSourceControlManagerQueries=true
+
+if [[ -d "$OUTPUT_ROOT/outputs" ]]; then
+  find "$OUTPUT_ROOT/outputs" -maxdepth 1 -name '*.json' -print0 | while IFS= read -r -d '' file; do
+    base="$(basename "$file")"
+    cp "$file" "$EDITABLE_DIR/$base"
+  done
+fi
+
+cat <<EOF
+
+Done.
+
+Generated files:
+- Latest parser output: $OUTPUT_ROOT/outputs
+- Editable copies:      $EDITABLE_DIR
+- Candidate fixtures:   $OUTPUT_ROOT/candidate-fixtures
+- Summary report:       $OUTPUT_ROOT/report.md
+- Machine index:        $OUTPUT_ROOT/index.json
+
+Recommended workflow:
+1. Edit files in: $EDITABLE_DIR
+2. Keep only the fields you really want the extractor to produce.
+3. Once reviewed, copy the corrected files into an approved fixtures directory.
+4. Re-run the benchmark and compare actual vs approved output.
+EOF