codingest 0.1.5__tar.gz → 0.1.7__tar.gz

{codingest-0.1.5 → codingest-0.1.7}/Cargo.lock

@@ -313,7 +313,7 @@ dependencies = [
 
 [[package]]
 name = "codingest"
-version = "0.1.5"
+version = "0.1.7"
 dependencies = [
  "aho-corasick",
  "base64",
@@ -345,7 +345,7 @@ dependencies = [
 
 [[package]]
 name = "codingest-cli"
-version = "0.1.5"
+version = "0.1.7"
 dependencies = [
  "anyhow",
  "clap",
@@ -357,7 +357,7 @@ dependencies = [
 
 [[package]]
 name = "codingest-mcp"
-version = "0.1.5"
+version = "0.1.7"
 dependencies = [
  "anyhow",
  "codingest",
@@ -366,7 +366,7 @@ dependencies = [
 
 [[package]]
 name = "codingest-py"
-version = "0.1.5"
+version = "0.1.7"
 dependencies = [
  "codingest",
  "codingest-cli",
@@ -1393,9 +1393,9 @@ dependencies = [
 
 [[package]]
 name = "kglite"
-version = "0.15.5"
+version = "0.15.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4664fc457e5dd762f5775977f06eb5377634556fffae45e4a86e2eb1e1eb0bfb"
+checksum = "a8cf3016156af6dafbc1041fa443e94002686b6deeee0fea38f9b712c50c96f8"
 dependencies = [
  "bzip2",
  "chrono",
@@ -1428,9 +1428,9 @@ dependencies = [
 
 [[package]]
 name = "kglite-mcp-server"
-version = "0.15.5"
+version = "0.15.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e5e5d2207a45eb77c90e60a772bc8cc60ceac3d7c6f3ad2b948e8079355e1df4"
+checksum = "afaf8e38626f3991d7d7fca93df9230ec7160af894e855bc5cbe49e6c0342ab3"
 dependencies = [
  "anyhow",
  "bytes",

{codingest-0.1.5 → codingest-0.1.7}/Cargo.toml

@@ -3,7 +3,7 @@ members = ["crates/codingest", "crates/codingest-cli", "crates/codingest-mcp", "
 resolver = "2"
 
 [workspace.package]
-version = "0.1.5"
+version = "0.1.7"
 edition = "2021"
 rust-version = "1.88"
 license = "MIT"
@@ -12,8 +12,12 @@ repository = "https://github.com/kkollsga/codingest"
 [workspace.dependencies]
 base64 = "0.22"
 sha2 = "0.11"
-# kglite engine + MCP server — imported as crates.io libraries. The 0.15.5
-# floor adds the two workspace controls codingest asked upstream for:
+# kglite engine + MCP server — imported as crates.io libraries. The 0.15.6
+# floor fixes mixed-selection vector dispatch, mixed-metric exact search,
+# community-detection modularity scoring, sampled-centrality validation, and
+# persisted HNSW validation, while adding the bulk node-degree primitive and
+# graph-algorithm/vector-search performance improvements. It builds on the two
+# workspace controls codingest asked upstream for in 0.15.5:
 # `workspace.sandbox_root`, a real containment boundary that confines
 # `set_root_dir` to a subtree instead of trusting the caller's path, and
 # `workspace.adopt_client_roots`. Both land on top of the generic
@@ -25,8 +29,8 @@ sha2 = "0.11"
 # and no disk storage. Neither patch changes `kglite::api`, graph output,
 # property encoding, or `.kgl` serialization.
 # Keep both libraries aligned so the graph and server share one engine release.
-kglite = { version = "0.15.5", default-features = false }
-kglite-mcp-server = { version = "0.15.5", default-features = false }
+kglite = { version = "0.15.6", default-features = false }
+kglite-mcp-server = { version = "0.15.6", default-features = false }
 
 # Mirror KGLite's release codegen so the standalone builder has
 # performance parity with the in-tree component (see KGLite

{codingest-0.1.5 → codingest-0.1.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: codingest
-Version: 0.1.5
+Version: 0.1.7
 Classifier: Development Status :: 3 - Alpha
 Classifier: Intended Audience :: Developers
 Classifier: Programming Language :: Python :: 3
@@ -12,7 +12,7 @@ Classifier: Programming Language :: Python :: 3.14
 Classifier: Programming Language :: Rust
 Classifier: Topic :: Software Development
 Classifier: Topic :: Software Development :: Libraries
-Requires-Dist: kglite>=0.15.5,<0.16
+Requires-Dist: kglite>=0.15.6,<0.16
 License-File: LICENSE
 Summary: Build fast, queryable code graphs for AI agents, MCP code review, and local or GitHub repository analysis.
 Keywords: ai-agents,mcp,code-review,knowledge-graph,code-analysis,github,tree-sitter,cypher
@@ -60,6 +60,10 @@ a real `kglite.KnowledgeGraph`, ready for Cypher queries.
 
 - **Agent-ready MCP:** give an MCP client `graph_overview`, `cypher_query`,
   `read_code_source`, repository switching, and automatic graph refresh.
+  [opencode is set up and verified end to end](https://codingest.readthedocs.io/en/latest/mcp.html#using-codingest-mcp-with-opencode)
+  — including a config that needs no absolute paths — and `codingest skill
+  install` packages the review skill for Claude Code and Codex, which opencode
+  also discovers with no extra step.
 - **Fast local review:** map definitions, callers, dependencies, routes,
   inheritance, and affected tests without uploading or executing the project.
 - **Open-source intelligence:** clone and analyse a GitHub repository with one
@@ -187,7 +191,7 @@ codingest skill install        # code-review Agent Skill
 ```
 
 The wheel bundles every grammar plus the `codingest` and `codingest-mcp`
-commands. KGLite ≥0.15.5 is installed automatically as the query/storage
+commands. KGLite ≥0.15.6 is installed automatically as the query/storage
 engine; its MCP server and the transitive `mcp-methods` framework power the
 builder-aware Codingest server. Nothing else needs to be installed.
 
@@ -200,7 +204,7 @@ Python prerequisites.
 ```toml
 [dependencies]
 codingest = "0.1"
-kglite = "0.15.5"
+kglite = "0.15.6"
 ```
 
 ```rust
@@ -272,7 +276,7 @@ diagnostic.
 ## Dependency policy
 
 `kglite` and `kglite-mcp-server` use matching crates.io requirements with a
-0.15.5 minimum and a shared lockfile. This keeps the builder, persistence
+0.15.6 minimum and a shared lockfile. This keeps the builder, persistence
 handoff, and embedded MCP server on one compatible engine patch line.
 
 ## Parity with the (now-removed) in-tree component

{codingest-0.1.5 → codingest-0.1.7}/README.md

@@ -33,6 +33,10 @@ a real `kglite.KnowledgeGraph`, ready for Cypher queries.
 
 - **Agent-ready MCP:** give an MCP client `graph_overview`, `cypher_query`,
   `read_code_source`, repository switching, and automatic graph refresh.
+  [opencode is set up and verified end to end](https://codingest.readthedocs.io/en/latest/mcp.html#using-codingest-mcp-with-opencode)
+  — including a config that needs no absolute paths — and `codingest skill
+  install` packages the review skill for Claude Code and Codex, which opencode
+  also discovers with no extra step.
 - **Fast local review:** map definitions, callers, dependencies, routes,
   inheritance, and affected tests without uploading or executing the project.
 - **Open-source intelligence:** clone and analyse a GitHub repository with one
@@ -160,7 +164,7 @@ codingest skill install        # code-review Agent Skill
 ```
 
 The wheel bundles every grammar plus the `codingest` and `codingest-mcp`
-commands. KGLite ≥0.15.5 is installed automatically as the query/storage
+commands. KGLite ≥0.15.6 is installed automatically as the query/storage
 engine; its MCP server and the transitive `mcp-methods` framework power the
 builder-aware Codingest server. Nothing else needs to be installed.
 
@@ -173,7 +177,7 @@ Python prerequisites.
 ```toml
 [dependencies]
 codingest = "0.1"
-kglite = "0.15.5"
+kglite = "0.15.6"
 ```
 
 ```rust
@@ -245,7 +249,7 @@ diagnostic.
 ## Dependency policy
 
 `kglite` and `kglite-mcp-server` use matching crates.io requirements with a
-0.15.5 minimum and a shared lockfile. This keeps the builder, persistence
+0.15.6 minimum and a shared lockfile. This keeps the builder, persistence
 handoff, and embedded MCP server on one compatible engine patch line.
 
 ## Parity with the (now-removed) in-tree component

{codingest-0.1.5 → codingest-0.1.7}/crates/codingest/Cargo.toml

@@ -10,7 +10,7 @@ keywords = ["knowledge-graph", "code-analysis", "tree-sitter", "cypher", "kglite
 categories = ["development-tools", "parser-implementations"]
 
 # Requires the kglite 0.15 engine API (`kglite::api::code_entities`); the
-# workspace pins the current 0.15 patch floor (0.15.5).
+# workspace pins the current 0.15 patch floor (0.15.6).
 
 [lib]
 name = "codingest"

{codingest-0.1.5 → codingest-0.1.7}/crates/codingest/src/bin/codingest_bench.rs

@@ -486,6 +486,11 @@ fn main() {
             "corpus_files": corpus.sha256.as_ref().map(|_| corpus.files),
             "corpus_bytes": corpus.sha256.as_ref().map(|_| corpus.bytes),
             "corpus_sha256": corpus.sha256,
+            // Fixed build configuration, echoed so a recorded row states it.
+            // Unlike codingest_stats, this harness has always built docs-ON,
+            // so the docs pass IS in its build times and node counts.
+            "include_tests": false,
+            "include_docs": true,
             "nodes": nodes,
             "edges": edges,
             "build_a_secs": (build_a_secs * 1000.0).round() / 1000.0,

{codingest-0.1.5 → codingest-0.1.7}/crates/codingest/src/bin/codingest_stats.rs

@@ -15,10 +15,21 @@
 //! Usage:
 //!   cargo run -p codingest --bin codingest_stats --release -- <path>
 //!   cargo run -p codingest --bin codingest_stats --release -- <path> --include-tests
+//!   cargo run -p codingest --bin codingest_stats --release -- <path> --include-docs
 //!   cargo run -p codingest --bin codingest_stats --release -- <path> --function-metrics
 //!   cargo run -p codingest --bin codingest_stats --release -- <path> --edge-breakdown
 //!   cargo run -p codingest --bin codingest_stats --release -- <path> --dump-calls a,b,c
 //!
+//! `--include-docs` turns the docs pass on. It is **off by default**, which is
+//! the configuration every historical row in `dev-docs/bench/results/` was
+//! measured at, so the default is the comparable one; without the flag `:Doc`,
+//! `:MENTIONS` and `:DOCUMENTS` are structurally absent and no docs-pass
+//! regression — in node counts or in build time — can be seen here. Both the
+//! tests and the docs configuration are echoed into the emitted JSON so a
+//! recorded row states the configuration it was taken under, and an unknown
+//! flag is a usage error rather than a silently-ignored typo that would report
+//! a measurement under a mode it never ran in.
+//!
 //! `--edge-breakdown` appends a per-connection-type edge histogram to the JSON
 //! object, with `IMPORTS` split by endpoint node type (`IMPORTS(File->File)`
 //! vs `IMPORTS(File->Module)`) — the two are the same connection type but
@@ -231,37 +242,76 @@ fn dump_calls(graph: &DirGraph, names: &BTreeSet<String>) -> Vec<CallRow> {
     rows
 }
 
+const USAGE: &str = "usage: codingest_stats <path> [--include-tests] [--include-docs] \
+     [--function-metrics] [--edge-breakdown] [--dump-calls name1,name2,...]";
+
+/// Everything the command line can say, parsed once so the configuration a run
+/// was taken under is a value that can be echoed and tested rather than a
+/// scatter of `any(|a| a == …)` predicates.
+#[derive(Debug, Default, PartialEq, Eq)]
+struct Options {
+    include_tests: bool,
+    include_docs: bool,
+    function_metrics: bool,
+    edge_breakdown: bool,
+    dump_calls: Option<BTreeSet<String>>,
+}
+
+/// Parse the arguments after `<path>`. Rejects anything unrecognised: a
+/// typo'd `--include-dcos` must not report a docs-off measurement as a docs-on
+/// one, which is the same reason `codingest_bench` rejects unknown flags.
+fn parse_options(args: &[String]) -> Result<Options, String> {
+    let mut out = Options::default();
+    let mut i = 0;
+    while i < args.len() {
+        match args[i].as_str() {
+            "--include-tests" => out.include_tests = true,
+            "--include-docs" => out.include_docs = true,
+            "--function-metrics" => out.function_metrics = true,
+            "--edge-breakdown" => out.edge_breakdown = true,
+            "--dump-calls" => {
+                let list = args.get(i + 1).ok_or_else(|| {
+                    "--dump-calls requires a comma-separated name list".to_string()
+                })?;
+                out.dump_calls = Some(
+                    list.split(',')
+                        .map(str::trim)
+                        .filter(|s| !s.is_empty())
+                        .map(str::to_string)
+                        .collect(),
+                );
+                i += 1;
+            }
+            other => return Err(format!("unknown argument `{other}`")),
+        }
+        i += 1;
+    }
+    Ok(out)
+}
+
 fn main() {
     let mut args = std::env::args().skip(1);
     let path = match args.next() {
         Some(p) => p,
         None => {
-            eprintln!(
-                "usage: codingest_stats <path> [--include-tests] [--function-metrics] \
-                 [--edge-breakdown] [--dump-calls name1,name2,...]"
-            );
+            eprintln!("{USAGE}");
             std::process::exit(2);
         }
     };
-    let options: Vec<_> = args.collect();
-    let include_tests = options.iter().any(|a| a == "--include-tests");
-    let emit_function_metrics = options.iter().any(|a| a == "--function-metrics");
-    let emit_edge_breakdown = options.iter().any(|a| a == "--edge-breakdown");
-    let dump_call_names: Option<BTreeSet<String>> = options
-        .iter()
-        .position(|a| a == "--dump-calls")
-        .map(|i| match options.get(i + 1) {
-            Some(list) => list
-                .split(',')
-                .map(str::trim)
-                .filter(|s| !s.is_empty())
-                .map(str::to_string)
-                .collect(),
-            None => {
-                eprintln!("--dump-calls requires a comma-separated name list");
-                std::process::exit(2);
-            }
-        });
+    let rest: Vec<String> = args.collect();
+    let options = match parse_options(&rest) {
+        Ok(options) => options,
+        Err(e) => {
+            eprintln!("{e}");
+            eprintln!("{USAGE}");
+            std::process::exit(2);
+        }
+    };
+    let include_tests = options.include_tests;
+    let include_docs = options.include_docs;
+    let emit_function_metrics = options.function_metrics;
+    let emit_edge_breakdown = options.edge_breakdown;
+    let dump_call_names = options.dump_calls.clone();
 
     let t = Instant::now();
     let (graph, stats) = match codingest::builder::run_with_options_stats(
@@ -270,7 +320,7 @@ fn main() {
         include_tests,
         None,
         None,
-        false,
+        include_docs,
     ) {
         Ok(pair) => pair,
         Err(e) => {
@@ -306,6 +356,7 @@ fn main() {
     let mut out = serde_json::json!({
         "path": path,
         "include_tests": include_tests,
+        "include_docs": include_docs,
         "build_secs": (build_secs * 1000.0).round() / 1000.0,
         "nodes": graph.graph.node_count(),
         "edges": graph.graph.edge_count(),
@@ -343,6 +394,97 @@ mod tests {
             .expect("corpus build")
     }
 
+    fn opts(args: &[&str]) -> Result<Options, String> {
+        parse_options(&args.iter().map(|a| a.to_string()).collect::<Vec<_>>())
+    }
+
+    #[test]
+    fn parse_options_defaults_to_the_historical_configuration() {
+        // Every row in dev-docs/bench/results/ predating --include-docs was
+        // measured tests-off/docs-off; the default must stay that way or old
+        // and new rows silently stop being comparable.
+        let parsed = opts(&[]).expect("empty argv parses");
+        assert_eq!(parsed, Options::default());
+        assert!(!parsed.include_tests);
+        assert!(!parsed.include_docs);
+    }
+
+    #[test]
+    fn parse_options_reads_every_flag_including_docs() {
+        let parsed = opts(&[
+            "--include-tests",
+            "--include-docs",
+            "--edge-breakdown",
+            "--function-metrics",
+            "--dump-calls",
+            "run, helper ,,other",
+        ])
+        .expect("all flags parse");
+        assert!(parsed.include_tests);
+        assert!(parsed.include_docs);
+        assert!(parsed.edge_breakdown);
+        assert!(parsed.function_metrics);
+        assert_eq!(
+            parsed.dump_calls,
+            Some(BTreeSet::from([
+                "run".to_string(),
+                "helper".to_string(),
+                "other".to_string()
+            ])),
+            "the name list is split, trimmed and empties dropped"
+        );
+    }
+
+    #[test]
+    fn parse_options_rejects_an_unknown_flag_but_not_a_dump_calls_value() {
+        // A typo must be a usage error: silently ignoring it would report a
+        // measurement under a configuration the run never used.
+        assert_eq!(
+            opts(&["--include-dcos"]),
+            Err("unknown argument `--include-dcos`".to_string())
+        );
+        assert_eq!(
+            opts(&["--dump-calls"]),
+            Err("--dump-calls requires a comma-separated name list".to_string())
+        );
+        // The value that follows --dump-calls is consumed, not re-inspected as
+        // a flag — even when it looks like one.
+        let parsed = opts(&["--dump-calls", "--include-docs", "--include-tests"])
+            .expect("the value after --dump-calls is a value");
+        assert!(!parsed.include_docs, "the value was not read as a flag");
+        assert!(parsed.include_tests);
+    }
+
+    #[test]
+    fn include_docs_is_what_makes_doc_nodes_visible() {
+        // The gap this flag closes: with the docs pass off — the hard-coded
+        // behaviour before it existed — a :Doc node cannot appear in any
+        // codingest_stats number, so no docs-pass regression can fail a gate.
+        fn doc_nodes(include_docs: bool) -> usize {
+            let graph = codingest::builder::run_with_options(
+                &corpus("docs_mdx"),
+                false,
+                true,
+                None,
+                None,
+                include_docs,
+            )
+            .expect("corpus build");
+            let params = HashMap::new();
+            let options = session::ExecuteOptions::eager(&params);
+            session::execute_read(&graph, "MATCH (d:Doc) RETURN d.name", &options)
+                .expect("query docs")
+                .result
+                .rows
+                .len()
+        }
+        assert_eq!(doc_nodes(false), 0, "docs off means no Doc node exists");
+        assert!(
+            doc_nodes(true) > 0,
+            "docs on must surface the docs_mdx corpus's Doc nodes"
+        );
+    }
+
     #[test]
     fn short_name_takes_the_rightmost_separator_of_any_flavour() {
         assert_eq!(short_name("pkg.util.helper"), "helper");