microservice-kg 0.1.0 → 0.1.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -1,7 +1,13 @@
1
1
  # microservice-kg
2
2
 
3
+ [![npm version](https://img.shields.io/npm/v/microservice-kg)](https://www.npmjs.com/package/microservice-kg)
4
+ [![GitHub repo](https://img.shields.io/badge/GitHub-pawarss1%2Fmicroservice--kg-black)](https://github.com/pawarss1/microservice-kg)
5
+
3
6
  `microservice-kg` builds a consolidated knowledge graph for a directory containing many services.
4
7
 
8
+ `npm`: [microservice-kg](https://www.npmjs.com/package/microservice-kg)
9
+ `GitHub`: [pawarss1/microservice-kg](https://github.com/pawarss1/microservice-kg)
10
+
5
11
  It is designed for multi-repo microservice workspaces where static code reasoning needs to happen at two levels at once:
6
12
 
7
13
  - service-to-service dependencies
@@ -17,22 +23,32 @@ The first version focuses on Java/Spring microservices and produces:
17
23
 
18
24
  ## Why this exists
19
25
 
20
- Tools like GitNexus are strong inside a single repo. This project adds a service-aware layer on top of a multi-repo workspace so you can answer:
26
+ In multi-service codebases, code search alone is not enough. You also need a durable model of:
27
+
28
+ - which services exist
29
+ - which services call which other services
30
+ - which class and method create that dependency
31
+ - which controller or provider method receives the traffic on the target side
32
+
33
+ `microservice-kg` adds that architecture layer on top of a multi-repo workspace so you can answer:
21
34
 
22
35
  - which services call which other services
23
36
  - which class and method create that dependency
24
37
  - which controller method receives the traffic on the target side
25
38
 
26
- That becomes valuable when a change crosses service boundaries and code search alone is not enough.
39
+ That becomes valuable when a change crosses service boundaries and a local file-by-file view is no longer sufficient.
27
40
 
28
41
  ## Why this is useful for agentic coding
29
42
 
30
- Agentic coding systems are good at editing code, but they often lack a durable architecture view across many services.
43
+ Public coding agents such as Claude Code and Codex are documented to inspect repositories, read and edit files, run commands, and gain extra context through MCP. Claude Code also documents a `ripgrep` dependency in its setup flow, which reinforces that file and search tooling are part of the normal local reasoning path.
44
+
45
+ That makes these systems strong at local code understanding, but they do not start with a durable relational model of service-to-service dependencies across a workspace.
31
46
 
32
47
  That creates a predictable failure mode:
33
48
 
34
49
  - the agent finds one caller and assumes it is the only caller
35
50
  - the agent edits one repo without seeing the upstream or downstream service contract
51
+ - the agent repeatedly reconstructs architecture from ad-hoc searches and partial code reads
36
52
  - the agent invents likely service relationships that are not actually present in code
37
53
 
38
54
  `microservice-kg` reduces that failure mode by turning service dependencies into queryable graph data.
@@ -63,6 +79,38 @@ with evidence instead of speculation.
63
79
 
64
80
  This does not eliminate hallucination entirely, but it narrows the space in which an agent has to infer behavior.
65
81
 
82
+ ## Screenshots
83
+
84
+ The images below are sanitized examples intended for the public README.
85
+
86
+ ### Inter-service graph
87
+
88
+ ![Inter-service dependency graph](docs/images/inter-service-example.png)
89
+
90
+ ### Cross-service neighborhood
91
+
92
+ ![Cross-service dependency neighborhood](docs/images/dependency-neighborhood-example.png)
93
+
94
+ ### Intra-service workflow path
95
+
96
+ ![Intra-service workflow path](docs/images/intra-service-workflow-example.png)
97
+
98
+ ### Obsidian graph view
99
+
100
+ ![Obsidian graph view](docs/images/obsidian-graph-view.png)
101
+
102
+ ## Install
103
+
104
+ ```bash
105
+ npm install -g microservice-kg
106
+ ```
107
+
108
+ Or run it without a global install:
109
+
110
+ ```bash
111
+ npx microservice-kg analyze /path/to/workspace --output ./output
112
+ ```
113
+
66
114
  ## Usage
67
115
 
68
116
  ```bash
@@ -159,3 +207,10 @@ Example questions an agent can ask through MCP:
159
207
  - gRPC
160
208
  - runtime reconciliation with OpenTelemetry service graphs
161
209
  - graph UI and interactive exploration
210
+
211
+ ## References
212
+
213
+ - [Claude Code settings](https://code.claude.com/docs/en/settings)
214
+ - [Claude Code setup](https://code.claude.com/docs/en/setup)
215
+ - [OpenAI Codex CLI](https://developers.openai.com/codex/cli)
216
+ - [How OpenAI uses Codex](https://openai.com/business/guides-and-resources/how-openai-uses-codex/)
package/package.json CHANGED
@@ -1,10 +1,14 @@
1
1
  {
2
2
  "name": "microservice-kg",
3
- "version": "0.1.0",
3
+ "version": "0.1.2",
4
4
  "description": "Consolidated microservice knowledge graph generator for multi-repo workspaces, with MCP support for agentic coding",
5
5
  "type": "module",
6
6
  "files": [
7
7
  "src",
8
+ "docs/images/inter-service-example.png",
9
+ "docs/images/dependency-neighborhood-example.png",
10
+ "docs/images/intra-service-workflow-example.png",
11
+ "docs/images/obsidian-graph-view.png",
8
12
  "README.md",
9
13
  "LICENSE"
10
14
  ],
package/src/analyzer.mjs CHANGED
@@ -6,7 +6,6 @@ const IGNORED_DIR_NAMES = new Set([
6
6
  ".idea",
7
7
  ".gradle",
8
8
  ".microservice-kg",
9
- ".gitnexus",
10
9
  ".claude",
11
10
  "build",
12
11
  "dist",