feat(cli): lectio --corpus-check validator

1 files changed, 69 insertions, 0 deletions

diff --git a/internal/cli/corpus.go b/internal/cli/corpus.go
new file mode 100644
index 0000000..15a140c
--- /dev/null
+++ b/internal/cli/corpus.go
@@ -0,0 +1,69 @@
+package cli
+
+import (
+	"encoding/json"
+	"fmt"
+	"io"
+	"path/filepath"
+	"strings"
+
+	"github.com/lukaszkasprzak/lectio/internal/bible"
+	"github.com/lukaszkasprzak/lectio/internal/config"
+)
+
+// runCorpusCheck handles --corpus-check <code|path>: validates a bible
+// corpus's sidecar + text (bible.CheckCorpus) and prints the report, plain or
+// (with asJSON) as JSON. It exits 1 if the report has any errors, 0
+// otherwise -- coverage warnings never fail the check.
+//
+// arg is either a bare corpus code, resolved against the user corpora dir
+// (config.CorporaDir(), which overrides the embedded built-ins the same way
+// normal rendering does), or a path to a <code>.tsv file -- its directory
+// becomes the user corpora dir for this run and its basename (minus ".tsv")
+// the code. The path form lets a corpus be validated before it is copied
+// into internal/bible/corpora/ (see scripts/corpus-validate.sh, make
+// add-corpus).
+func runCorpusCheck(arg string, asJSON bool, stdout, stderr io.Writer) int {
+	code := arg
+	if strings.Contains(arg, "/") || strings.HasSuffix(arg, ".tsv") {
+		dir := filepath.Dir(arg)
+		code = strings.TrimSuffix(filepath.Base(arg), ".tsv")
+		bible.SetUserCorporaDir(dir)
+	} else if dir, err := config.CorporaDir(); err == nil {
+		bible.SetUserCorporaDir(dir)
+	}
+
+	report := bible.CheckCorpus(code)
+
+	if asJSON {
+		enc := json.NewEncoder(stdout)
+		enc.SetIndent("", "  ")
+		if err := enc.Encode(report); err != nil {
+			fmt.Fprintln(stderr, "lectio:", err)
+			return 1
+		}
+	} else {
+		printCorpusReport(stdout, report)
+	}
+
+	if !report.OK() {
+		return 1
+	}
+	return 0
+}
+
+// printCorpusReport renders a CorpusReport as plain text: one line per
+// error/warning, then a summary line.
+func printCorpusReport(w io.Writer, r bible.CorpusReport) {
+	for _, e := range r.Errors {
+		fmt.Fprintf(w, "ERROR: %s\n", e)
+	}
+	for _, wm := range r.Warnings {
+		fmt.Fprintf(w, "WARNING: %s\n", wm)
+	}
+	if r.OK() {
+		fmt.Fprintf(w, "ok: %s (%d warning(s))\n", r.Code, len(r.Warnings))
+	} else {
+		fmt.Fprintf(w, "FAIL: %s (%d error(s), %d warning(s))\n", r.Code, len(r.Errors), len(r.Warnings))
+	}
+}