|
| 1 | +#!/usr/bin/env bash |
| 2 | +set -euo pipefail |
| 3 | + |
| 4 | +# Index a single benchmark repository and capture metrics. |
| 5 | +# Usage: benchmark-index.sh <binary> <lang> <repo_path> <results_dir> |
| 6 | + |
| 7 | +BINARY="${1:?Usage: benchmark-index.sh <binary> <lang> <repo_path> <results_dir>}" |
| 8 | +LANG="${2:?}" |
| 9 | +REPO="${3:?}" |
| 10 | +RESULTS_DIR="${4:?}" |
| 11 | + |
| 12 | +# Resolve symlinks |
| 13 | +REPO=$(cd "$REPO" && pwd -P) |
| 14 | + |
| 15 | +OUT="$RESULTS_DIR/$LANG" |
| 16 | +mkdir -p "$OUT" |
| 17 | + |
| 18 | +echo "INDEX: $LANG ($REPO)" |
| 19 | + |
| 20 | +# Count source files and LOC (exclude .git, vendor, node_modules, build dirs) |
| 21 | +FILE_COUNT=$(find "$REPO" -type f \ |
| 22 | + ! -path '*/.git/*' ! -path '*/node_modules/*' ! -path '*/vendor/*' \ |
| 23 | + ! -path '*/target/*' ! -path '*/build/*' ! -path '*/dist/*' \ |
| 24 | + ! -path '*/__pycache__/*' ! -path '*/.cache/*' \ |
| 25 | + | wc -l | tr -d ' ') |
| 26 | + |
| 27 | +LOC=$(find "$REPO" -type f \ |
| 28 | + ! -path '*/.git/*' ! -path '*/node_modules/*' ! -path '*/vendor/*' \ |
| 29 | + ! -path '*/target/*' ! -path '*/build/*' ! -path '*/dist/*' \ |
| 30 | + ! -path '*/__pycache__/*' ! -path '*/.cache/*' \ |
| 31 | + -exec cat {} + 2>/dev/null | wc -l | tr -d ' ') |
| 32 | + |
| 33 | +echo "$FILE_COUNT" > "$OUT/file-count.txt" |
| 34 | +echo "$LOC" > "$OUT/loc.txt" |
| 35 | + |
| 36 | +# Index via CLI and capture timing |
| 37 | +START_MS=$(python3 -c "import time; print(int(time.time()*1000))") |
| 38 | + |
| 39 | +INDEX_JSON=$("$BINARY" cli index_repository "{\"repo_path\":\"$REPO\",\"mode\":\"full\"}" 2>/dev/null || echo '{"error":"index failed"}') |
| 40 | + |
| 41 | +END_MS=$(python3 -c "import time; print(int(time.time()*1000))") |
| 42 | +ELAPSED=$((END_MS - START_MS)) |
| 43 | + |
| 44 | +echo "$INDEX_JSON" > "$OUT/00-index.json" |
| 45 | +echo "$ELAPSED" > "$OUT/index-time.txt" |
| 46 | + |
| 47 | +# Extract node/edge counts (CLI wraps in MCP content envelope) |
| 48 | +NODES=$(echo "$INDEX_JSON" | python3 -c " |
| 49 | +import json,sys |
| 50 | +d=json.load(sys.stdin) |
| 51 | +# Unwrap MCP content envelope if present |
| 52 | +if 'content' in d: |
| 53 | + inner=json.loads(d['content'][0]['text']) |
| 54 | +else: |
| 55 | + inner=d |
| 56 | +print(inner.get('nodes',0)) |
| 57 | +" 2>/dev/null || echo "0") |
| 58 | +EDGES=$(echo "$INDEX_JSON" | python3 -c " |
| 59 | +import json,sys |
| 60 | +d=json.load(sys.stdin) |
| 61 | +if 'content' in d: |
| 62 | + inner=json.loads(d['content'][0]['text']) |
| 63 | +else: |
| 64 | + inner=d |
| 65 | +print(inner.get('edges',0)) |
| 66 | +" 2>/dev/null || echo "0") |
| 67 | +PROJECT=$(echo "$INDEX_JSON" | python3 -c " |
| 68 | +import json,sys |
| 69 | +d=json.load(sys.stdin) |
| 70 | +if 'content' in d: |
| 71 | + inner=json.loads(d['content'][0]['text']) |
| 72 | +else: |
| 73 | + inner=d |
| 74 | +print(inner.get('project','')) |
| 75 | +" 2>/dev/null || echo "") |
| 76 | + |
| 77 | +echo "$NODES" > "$OUT/nodes.txt" |
| 78 | +echo "$EDGES" > "$OUT/edges.txt" |
| 79 | +echo "$PROJECT" > "$OUT/project.txt" |
| 80 | + |
| 81 | +printf " %s: %s files, %s LOC, %sms, %s nodes, %s edges\n" \ |
| 82 | + "$LANG" "$FILE_COUNT" "$LOC" "$ELAPSED" "$NODES" "$EDGES" |
0 commit comments