recoco

This is the core package for ReCoco, core provides direct access to ReCoco's complete functionality. ReCoco is a rust-only fork of CocoIndex

When to use this crate

Use recoco when you want:

✅ To use multiple ReCoco components together
✅ Feature parity with the full ReCoco ecosystem
✅ Easy dependency management

Use individual crates (recoco-core, recoco-utils, recoco-splitters) when you want:

⚡ Fine-grained dependency control
⚡ Minimal compile times
⚡ Only specific functionality (e.g., just utils)

Installation

[dependencies]
recoco = { version = "0.2", features = ["function-split", "source-postgres"] }

📦 Feature Flags

This crate mirrors all features from recoco-core. Enable only what you need to keep dependencies minimal.

🎯 Default Features

recoco = "0.2"  # Includes: persistence, server, source-local-file

📦 Feature Bundles (Convenience)

Feature	Description	Use When
`full`	Everything (⚠️ very heavy)	You need all functionality
`all-sources`	All data source connectors	Working with multiple data sources
`all-targets`	All data target connectors	Writing to multiple databases
`all-functions`	All transformation functions	Need all data processing capabilities
`all-llm-providers`	All LLM provider integrations	Working with multiple AI APIs
`all-splitter-languages`	All Tree-sitter grammars	Processing many programming languages

📥 Sources (Data Ingestion)

Feature	Description
`source-local-file`	Local filesystem (✅ default)
`source-postgres`	PostgreSQL with CDC
`source-s3`	Amazon S3
`source-azure`	Azure Blob Storage
`source-gdrive`	Google Drive

📤 Targets (Data Persistence)

Feature	Description
`target-postgres`	PostgreSQL database
`target-qdrant`	Qdrant vector database
`target-neo4j`	Neo4j graph database
`target-kuzu`	Kùzu embedded graph database

⚙️ Functions (Data Transformations)

Feature	Description
`function-split`	Text splitting (recursive, semantic)
`function-embed`	Generate text embeddings
`function-extract-llm`	LLM-based data extraction
`function-detect-lang`	Programming language detection
`function-json`	JSON/JSON5 parsing

🤖 LLM Providers

Feature	Provider
`provider-openai`	OpenAI (GPT-4, etc.)
`provider-anthropic`	Anthropic (Claude)
`provider-azure`	Azure OpenAI
`provider-gemini`	Google Gemini
`provider-bedrock`	AWS Bedrock
`provider-ollama`	Ollama (local LLMs)
`provider-voyage`	Voyage AI (embeddings)
`provider-litellm`	LiteLLM (unified gateway)
`provider-openrouter`	OpenRouter (multi-provider)
`provider-vllm`	vLLM (inference server)

🔤 Splitter Languages (Tree-sitter Grammars)

Enable specific programming language support for code splitting:

Feature	Language
`splitter-language-c`	C
`splitter-language-c-sharp`	C#
`splitter-language-cpp`	C++
`splitter-language-css`	CSS
`splitter-language-fortran`	Fortran
`splitter-language-go`	Go
`splitter-language-html`	HTML
`splitter-language-java`	Java
`splitter-language-javascript`	JavaScript
`splitter-language-json`	JSON
`splitter-language-kotlin`	Kotlin
`splitter-language-markdown`	Markdown
`splitter-language-php`	PHP
`splitter-language-python`	Python
`splitter-language-r`	R
`splitter-language-ruby`	Ruby
`splitter-language-rust`	Rust
`splitter-language-scala`	Scala
`splitter-language-sql`	SQL
`splitter-language-swift`	Swift
`splitter-language-toml`	TOML
`splitter-language-typescript`	TypeScript
`splitter-language-xml`	XML
`splitter-language-yaml`	YAML

🏗️ Core Features

Feature	Description
`persistence`	Database-backed state tracking (✅ default)
`server`	HTTP server components (✅ default)
`json-schema`	JSON Schema support

🎯 Common Use Cases

Local File Processing

recoco = { version = "0.2", default-features = false, features = [
    "source-local-file",
    "function-split",
    "splitter-language-rust"
]}

RAG Pipeline with OpenAI

recoco = { version = "0.2", features = [
    "source-s3",
    "function-split",
    "function-embed",
    "provider-openai",
    "target-qdrant",
    "all-splitter-languages"
]}

Database ETL

recoco = { version = "0.2", features = [
    "source-postgres",
    "target-postgres",
    "function-json"
]}

Multi-Cloud Data Sync

recoco = { version = "0.2", features = [
    "all-sources",
    "all-targets",
    "batching"
]}

Lightweight Transient Processing (No Database)

recoco = { version = "0.2", default-features = false, features = [
    "function-split",
    "function-json"
]}

📚 Documentation

Main README: ../../README.md
API Docs: docs.rs/recoco
Examples: examples/
recoco-utils: ../recoco-utils/README.md

🔧 Development

This crate is part of the ReCoco workspace:

# Build with specific features
cargo build -p recoco --features "function-split,source-postgres"

# Test with all features
cargo test -p recoco --features full

# Run examples
cargo run -p recoco --example transient --features function-split

📄 License

Apache-2.0. See main repository for details.

recoco-core 0.2.1