#!/bin/bash

# VisionFlow Codebase Blob Assembler

# Assembles client and server code into concatenated text files

# for 1M token context window audit

#

# Output: /tmp/visionflow-blobs/

# - server-blob.txt (~560K tokens)

# - client-blob.txt (~306K tokens)

# - spec-blob.txt (~83K tokens)

# - manifest.json (metadata index)

set -uo pipefail

PROJECT="/home/devuser/workspace/project"

OUTPUT="/tmp/visionflow-blobs"

mkdir -p "$OUTPUT"

# ============================================================

# UTILITY: Add file with metadata header

# ============================================================

add_file() {

local blob_file="$1"

local src_file="$2"

local category="${3:-source}"

if [ ! -f "$src_file" ]; then

return

fi

local rel_path="${src_file#$PROJECT/}"

local lines=$(wc -l < "$src_file")

echo "" >> "$blob_file"

echo "// ================================================================" >> "$blob_file"

echo "// FILE: $rel_path" >> "$blob_file"

echo "// LINES: $lines | CATEGORY: $category" >> "$blob_file"

echo "// ================================================================" >> "$blob_file"

cat "$src_file" >> "$blob_file"

echo "" >> "$blob_file"

}

# ============================================================

# UTILITY: Strip #[cfg(test)] blocks from Rust files

# ============================================================

add_rust_file_no_tests() {

local blob_file="$1"

local src_file="$2"

local category="${3:-source}"

if [ ! -f "$src_file" ]; then

return

fi

local rel_path="${src_file#$PROJECT/}"

# Strip cfg(test) modules using awk

local content

content=$(awk '

BEGIN { skip=0; brace_count=0 }

/^#\[cfg\(test\)\]/ { skip=1; next }

skip==1 && /\{/ { brace_count++; next }

skip==1 && /\}/ { brace_count--; if(brace_count<=0) { skip=0; brace_count=0 }; next }

skip==1 { next }

{ print }

' "$src_file")

local lines=$(echo "$content" | wc -l)

echo "" >> "$blob_file"

echo "// ================================================================" >> "$blob_file"

echo "// FILE: $rel_path" >> "$blob_file"

echo "// LINES: $lines | CATEGORY: $category (tests stripped)" >> "$blob_file"

echo "// ================================================================" >> "$blob_file"

echo "$content" >> "$blob_file"

echo "" >> "$blob_file"

}

# ============================================================

# SERVER BLOB

# ============================================================

echo "=== Assembling Server Blob ==="

SERVER_BLOB="$OUTPUT/server-blob.txt"

cat > "$SERVER_BLOB" << 'HEADER'

================================================================================

VISIONFLOW SERVER CODEBASE (Rust/Actix-web)

================================================================================

Architecture: Actix-web + Actor system + CQRS + Hexagonal + GPU (CUDA)

Database: Neo4j graph database

Auth: Nostr NIP-98 + session tokens

Physics: GPU-accelerated force-directed layout with CUDA kernels

Features: OWL ontology reasoning, semantic analysis, graph algorithms,

community detection, anomaly detection, pathfinding, stress majorization

Total files: ~427 Rust source files

================================================================================

HEADER

# 1. Entry point and app state

echo " [1/12] Entry points..."

add_rust_file_no_tests "$SERVER_BLOB" "$PROJECT/src/main.rs" "entry"

add_rust_file_no_tests "$SERVER_BLOB" "$PROJECT/src/lib.rs" "entry"

add_rust_file_no_tests "$SERVER_BLOB" "$PROJECT/src/app_state.rs" "entry"

add_rust_file_no_tests "$SERVER_BLOB" "$PROJECT/src/openapi.rs" "entry"

# 2. Models

echo " [2/12] Models..."

find "$PROJECT/src/models/" -name "*.rs" -type f 2>/dev/null | sort | while read f; do

add_rust_file_no_tests "$SERVER_BLOB" "$f" "model"

done

# 3. Ports (interfaces)

echo " [3/12] Ports..."

find "$PROJECT/src/ports/" -name "*.rs" -type f 2>/dev/null | sort | while read f; do

add_rust_file_no_tests "$SERVER_BLOB" "$f" "port"

done

# 4. Errors

echo " [4/12] Errors..."

find "$PROJECT/src/errors/" -name "*.rs" -type f 2>/dev/null | sort | while read f; do

add_rust_file_no_tests "$SERVER_BLOB" "$f" "error"

done

# 5. Adapters (excluding test files)

echo " [5/12] Adapters..."

find "$PROJECT/src/adapters/" -name "*.rs" -type f ! -path "*/tests/*" 2>/dev/null | sort | while read f; do

add_rust_file_no_tests "$SERVER_BLOB" "$f" "adapter"

done

# 6. CQRS layer

echo " [6/12] CQRS..."

find "$PROJECT/src/cqrs/" -name "*.rs" -type f 2>/dev/null | sort | while read f; do

add_rust_file_no_tests "$SERVER_BLOB" "$f" "cqrs"

done

# 7. Application services

echo " [7/12] Application..."

find "$PROJECT/src/application/" -name "*.rs" -type f ! -path "*/tests/*" 2>/dev/null | sort | while read f; do

add_rust_file_no_tests "$SERVER_BLOB" "$f" "application"

done

# 8. Services

echo " [8/12] Services..."

find "$PROJECT/src/services/" -name "*.rs" -type f ! -name "*stub*" 2>/dev/null | sort | while read f; do

add_rust_file_no_tests "$SERVER_BLOB" "$f" "service"

done

# 9. Actors

echo " [9/12] Actors..."

find "$PROJECT/src/actors/" -name "*.rs" -type f 2>/dev/null | sort | while read f; do

add_rust_file_no_tests "$SERVER_BLOB" "$f" "actor"

done

# 10. Handlers

echo " [10/12] Handlers..."

find "$PROJECT/src/handlers/" -name "*.rs" -type f ! -path "*/tests/*" 2>/dev/null | sort | while read f; do

add_rust_file_no_tests "$SERVER_BLOB" "$f" "handler"

done

# 11. GPU + Physics

echo " [11/12] GPU & Physics..."

find "$PROJECT/src/gpu/" -name "*.rs" -type f 2>/dev/null | sort | while read f; do

add_rust_file_no_tests "$SERVER_BLOB" "$f" "gpu"

done

find "$PROJECT/src/physics/" -name "*.rs" -type f ! -name "*integration_test*" 2>/dev/null | sort | while read f; do

add_rust_file_no_tests "$SERVER_BLOB" "$f" "physics"

done

# 12. Infrastructure (config, middleware, utils, events, constraints, types, etc)

echo " [12/12] Infrastructure..."

for dir in config middleware utils events constraints types inference ontology telemetry protocols settings validation; do

find "$PROJECT/src/$dir/" -name "*.rs" -type f ! -path "*/tests/*" ! -name "*test*" 2>/dev/null | sort | while read f; do

add_rust_file_no_tests "$SERVER_BLOB" "$f" "$dir"

done 2>/dev/null || true

done

# Add Cargo.toml for dependency context

echo "" >> "$SERVER_BLOB"

echo "// ================================================================" >> "$SERVER_BLOB"

echo "// FILE: Cargo.toml" >> "$SERVER_BLOB"

echo "// CATEGORY: build-config" >> "$SERVER_BLOB"

echo "// ================================================================" >> "$SERVER_BLOB"

cat "$PROJECT/Cargo.toml" >> "$SERVER_BLOB"

SERVER_LINES=$(wc -l < "$SERVER_BLOB")

echo " Server blob: $SERVER_LINES lines"

# ============================================================

# CLIENT BLOB

# ============================================================

echo ""

echo "=== Assembling Client Blob ==="

CLIENT_BLOB="$OUTPUT/client-blob.txt"

cat > "$CLIENT_BLOB" << 'HEADER'

================================================================================

VISIONFLOW CLIENT CODEBASE (TypeScript/React/Three.js)

================================================================================

Framework: React 19 + Three.js/R3F + Zustand + TailwindCSS 4

3D: @react-three/fiber + @react-three/drei + @react-three/xr

WASM: scene-effects crate (Rust compiled to WebAssembly)

State: Zustand stores with Immer

UI: Radix UI primitives + Lucide icons + Framer Motion

WebSocket: Binary protocol for real-time position updates

Features: 3D graph visualization, XR/VR mode, voice commands,

AI integrations (RAGFlow, Perplexity, OpenAI, Kokoro TTS, Whisper STT)

================================================================================

HEADER

# Client files by feature area (excluding tests)

echo " [1/7] Core app files..."

find "$PROJECT/client/src" -maxdepth 1 \( -name "*.ts" -o -name "*.tsx" \) 2>/dev/null | sort | while read f; do

add_file "$CLIENT_BLOB" "$f" "core"

done

echo " [2/7] Features (graph, settings, etc)..."

find "$PROJECT/client/src/features/" \( -name "*.ts" -o -name "*.tsx" \) 2>/dev/null | grep -v "__tests__" | grep -v ".test." | grep -v ".spec." | sort | while read f; do

add_file "$CLIENT_BLOB" "$f" "feature"

done

echo " [3/7] Components..."

find "$PROJECT/client/src/components/" \( -name "*.ts" -o -name "*.tsx" \) 2>/dev/null | grep -v "__tests__" | grep -v ".test." | grep -v ".spec." | sort | while read f; do

add_file "$CLIENT_BLOB" "$f" "component"

done

echo " [4/7] Hooks & Store..."

find "$PROJECT/client/src/hooks/" "$PROJECT/client/src/store/" \( -name "*.ts" -o -name "*.tsx" \) 2>/dev/null | grep -v "__tests__" | grep -v ".test." | sort | while read f; do

add_file "$CLIENT_BLOB" "$f" "hook-store"

done

echo " [5/7] Services & API..."

find "$PROJECT/client/src/services/" "$PROJECT/client/src/api/" -name "*.ts" 2>/dev/null | grep -v "__tests__" | grep -v ".test." | sort | while read f; do

add_file "$CLIENT_BLOB" "$f" "service"

done

echo " [6/7] Types, Utils, Config, Settings..."

for dir in types utils config settings; do

if [ -d "$PROJECT/client/src/$dir/" ]; then

find "$PROJECT/client/src/$dir/" -name "*.ts" 2>/dev/null | grep -v "__tests__" | grep -v ".test." | sort | while read f; do

add_file "$CLIENT_BLOB" "$f" "types-utils"

done

fi

done

echo " [7/7] Workers, WASM bridge..."

find "$PROJECT/client/src/workers/" "$PROJECT/client/src/wasm/" -name "*.ts" 2>/dev/null | grep -v ".test." | sort | while read f; do

add_file "$CLIENT_BLOB" "$f" "worker-wasm"

done

# WASM Rust crate

echo "" >> "$CLIENT_BLOB"

echo "// ================================================================" >> "$CLIENT_BLOB"

echo "// WASM SCENE EFFECTS CRATE (Rust -> WebAssembly)" >> "$CLIENT_BLOB"

echo "// ================================================================" >> "$CLIENT_BLOB"

find "$PROJECT/client/crates/" -name "*.rs" 2>/dev/null | sort | while read f; do

add_file "$CLIENT_BLOB" "$f" "wasm-rust"

done

# Package.json for dependency context

echo "" >> "$CLIENT_BLOB"

echo "// ================================================================" >> "$CLIENT_BLOB"

echo "// FILE: client/package.json" >> "$CLIENT_BLOB"

echo "// CATEGORY: build-config" >> "$CLIENT_BLOB"

echo "// ================================================================" >> "$CLIENT_BLOB"

cat "$PROJECT/client/package.json" >> "$CLIENT_BLOB"

CLIENT_LINES=$(wc -l < "$CLIENT_BLOB")

echo " Client blob: $CLIENT_LINES lines"

# ============================================================

# SPEC BLOB (Diagrams + Dense Overview)

# ============================================================

echo ""

echo "=== Assembling Spec Blob ==="

SPEC_BLOB="$OUTPUT/spec-blob.txt"

cat > "$SPEC_BLOB" << 'HEADER'

================================================================================

VISIONFLOW FUNCTIONAL SPECIFICATION & ARCHITECTURE DIAGRAMS

================================================================================

Contains: Dense system overview, Mermaid architecture diagrams,

data flow diagrams, component diagrams, API reference

All diagrams are Mermaid notation (diagrams-as-code)

================================================================================

HEADER

# Dense overview first (most valuable single doc)

echo " [1/4] Dense overview..."

add_file "$SPEC_BLOB" "$PROJECT/docs/denseOverview.md" "spec"

# Architecture overview

echo " [2/4] Architecture docs..."

add_file "$SPEC_BLOB" "$PROJECT/docs/architecture.md" "spec"

add_file "$SPEC_BLOB" "$PROJECT/docs/api-reference.md" "spec"

# Dedicated diagram files (high-density architecture diagrams)

echo " [3/4] Architecture diagrams..."

find "$PROJECT/docs/diagrams/" -name "*.md" -type f 2>/dev/null | sort | while read f; do

add_file "$SPEC_BLOB" "$f" "diagram"

done

# Key architecture explanation docs

echo " [4/4] Architecture explanations..."

for f in \

"$PROJECT/docs/explanation/system-overview.md" \

"$PROJECT/docs/explanation/architecture/data-flow.md" \

"$PROJECT/docs/explanation/architecture/system-architecture.md" \

"$PROJECT/docs/explanation/architecture/server/overview.md" \

"$PROJECT/docs/explanation/architecture/client/overview.md" \

"$PROJECT/docs/explanation/architecture/components/websocket-protocol.md" \

"$PROJECT/docs/explanation/architecture/gpu/communication-flow.md" \

"$PROJECT/docs/explanation/architecture/event-driven-architecture.md" \

"$PROJECT/docs/explanation/architecture/pipeline-sequence-diagrams.md" \

"$PROJECT/docs/explanation/concepts/architecture/core/server.md" \

"$PROJECT/docs/explanation/concepts/architecture/core/client.md" \

"$PROJECT/docs/explanation/concepts/actor-model.md" \

"$PROJECT/docs/explanation/concepts/physics-engine.md" \

"$PROJECT/docs/explanation/concepts/hexagonal-architecture.md" \

"$PROJECT/docs/reference/protocols/binary-websocket.md" \

"$PROJECT/docs/reference/api/websocket-endpoints.md" \

"$PROJECT/docs/reference/database/ontology-schema-v2.md" \

"$PROJECT/docs/reference/database/schemas.md" \

; do

add_file "$SPEC_BLOB" "$f" "architecture-doc" 2>/dev/null || true

done

SPEC_LINES=$(wc -l < "$SPEC_BLOB")

echo " Spec blob: $SPEC_LINES lines"

# ============================================================

# MANIFEST

# ============================================================

echo ""

echo "=== Generating Manifest ==="

cat > "$OUTPUT/manifest.json" << EOF

{

"project": "VisionFlow AI Multi-Agent Knowledge Graph Visualisation",

"generated": "$(date -u +%Y-%m-%dT%H:%M:%SZ)",

"blobs": {

"server-blob.txt": {

"lines": $SERVER_LINES,

"estimated_tokens": $(echo "$SERVER_LINES * 3.5 / 1" | bc),

"language": "Rust",

"framework": "Actix-web + CUDA",

"contents": "All server source code (tests stripped), Cargo.toml"

},

"client-blob.txt": {

"lines": $CLIENT_LINES,

"estimated_tokens": $(echo "$CLIENT_LINES * 3.5 / 1" | bc),

"language": "TypeScript/TSX",

"framework": "React 19 + Three.js + R3F",

"contents": "All client source code (tests excluded), WASM crate, package.json"

},

"spec-blob.txt": {

"lines": $SPEC_LINES,

"estimated_tokens": $(echo "$SPEC_LINES * 2.5 / 1" | bc),

"language": "Markdown + Mermaid",

"contents": "Dense overview, architecture diagrams, API reference, key docs"

}

},

"total_estimated_tokens": $(echo "($SERVER_LINES * 3.5 + $CLIENT_LINES * 3.5 + $SPEC_LINES * 2.5) / 1" | bc),

"target_context_window": 1000000

}

EOF

echo ""

echo "=== DONE ==="

echo "Files in $OUTPUT/:"

ls -lh "$OUTPUT/"

echo ""

echo "Token estimates:"

cat "$OUTPUT/manifest.json" | python3 -m json.tool 2>/dev/null || cat "$OUTPUT/manifest.json"