How do their APIs compare?

I want to be precise here rather than force a comparison that doesn't really apply — these two repos aren't API-comparable in a meaningful way, because they serve entirely different purposes:

codingapi/tx-lcn

This is a distributed transaction framework, and its "API" surface is about integration points for transactional coordination across services:

• Auto-configuration-driven Spring integration: modules like TCAutoConfiguration in [txlcn-tc/src/main/java/com/codingapi/txlcn/tc/TCAutoConfiguration.java:L1-L22] expose annotation-based configuration (@Configuration) rather than a hand-called API — consumers add a starter dependency (e.g. starter-txlcn-tc, per the README's code structure section) and the framework wires itself in.
• JDBC event-listener SPI: the txlcn-p6spy module (adapted from the P6Spy project) exposes an extensible listener API via JdbcEventListener [txlcn-p6spy/src/main/java/com/codingapi/txlcn/p6spy/event/JdbcEventListener.java:L26-L38], which is intentionally an abstract class rather than an interface so implementations aren't broken as methods are added. Listeners can be registered via a CompoundJdbcEventListener [txlcn-p6spy/src/main/java/com/codingapi/txlcn/p6spy/CompoundJdbcEventListener.java:L1-L30] which composes a DefaultEventListener with custom P6spyJdbcEventListener implementations.
• The overall architecture separates a TC (transaction client) module and TM (transaction manager) module, per the README's code structure, communicating over a defined protocol (txlcn-protocol).
• The RepoPilot verdict for this repo is GO — healthy across all four use cases, suggesting the architecture and dependency signals are sound, though no specific cycle/dependency counts were surfaced in the excerpt I have.

crossoverjie/jcsprout

This repo is not a library with a public API at all — it's a personal collection of Java learning exercises and data-structure/algorithm implementations. What look like "APIs" here are just standalone classes with main methods for demonstration/testing, e.g.:

• Search — a simple in-memory word-search/autocomplete demo using a ConcurrentHashMap, with pattern matching logic in search(String keyWord) [src/main/java/com/crossoverjie/actual/Search.java:L141-L158].
• LRUAbstractMap — a hand-rolled LRU cache implementation with size/timeout eviction callbacks, described in its class comment [src/main/java/com/crossoverjie/actual/LRUAbstractMap.java:L11-L18].
• BinaryNode / BinaryNodeTravel — binary tree node classes with getters/setters and level-order traversal methods like levelIterator [src/main/java/com/crossoverjie/algorithm/BinaryNode.java:L67-L90].
• BloomFilters — a bloom filter demo with multiple hash functions [src/main/java/com/crossoverjie/algorithm/BloomFilters.java:L50-L89].

Bottom line: tx-lcn has a real integration surface (Spring auto-configuration + JDBC event SPI + inter-service transaction protocol) meant for production use in distributed systems. jcsprout has no consumable API — it's educational/reference code with public static void main entry points for individual concepts. If you're trying to integrate one of these into a project, only tx-lcn is designed for that; jcsprout is meant to be read, not depended on.

FOLLOW_UPS:

• How does tx-lcn's TC and TM communicate over its protocol?
• What does the p6spy module do in tx-lcn?
• Are there usage examples for LRUAbstractMap in jcsprout?