Resolve pipeline settings and collection selections

internal/cli/process_command.go

@@ -1,6 +1,7 @@
 package cli
 
 import (
+	"fmt"
 	"strconv"
 
 	commandkit "go.maximumdirect.net/commandkit"
@@ -73,6 +74,20 @@ func processCommand() commandkit.Command {
 				}
 			}))
 	}
+	for _, item := range []struct {
+		name string
+		path ck.Path
+	}{
+		{"glossaries", ck.Keys("glossaries")}, {"protected-terms", ck.Keys("protected_terms")},
+	} {
+		name, path := item.name, item.path
+		flags = append(flags, commandkit.ValueFlag(commandkit.FlagMeta{Name: name, Usage: "JSON array of collection names, or null."}, (*any)(nil), parseCollectionNames,
+			func(in *processInput, value commandkit.Supplied[any]) {
+				if value.Set {
+					in.Overrides = append(in.Overrides, ck.Override{Path: path, Value: value.Value, Flag: "--" + name})
+				}
+			}))
+	}
 	flags = append(flags, commandkit.BoolFlag(commandkit.FlagMeta{Name: "reporting", Usage: "Enable ordinary reports."}, false,
 		func(in *processInput, v commandkit.Supplied[bool]) {
 			if v.Set {
@@ -83,3 +98,23 @@ func processCommand() commandkit.Command {
 		Arguments: []commandkit.Argument[processInput]{commandkit.ValueArgument(commandkit.ArgumentMeta{Name: "transcript", Usage: "Transcript JSON file."}, true, stringValue, func(in *processInput, v commandkit.Supplied[string]) { in.Transcript = v.Value })},
 	})
 }
+
+func parseCollectionNames(raw string) (any, error) {
+	value, err := commandkit.ParseCLIJSON(raw)
+	if err != nil {
+		return nil, err
+	}
+	if value == nil {
+		return nil, nil
+	}
+	names, ok := value.([]any)
+	if !ok {
+		return nil, fmt.Errorf("expected a JSON array of collection names or null")
+	}
+	for _, name := range names {
+		if _, ok := name.(string); !ok {
+			return nil, fmt.Errorf("collection names must be strings")
+		}
+	}
+	return names, nil
+}

internal/cli/process_command_integration_test.go

@@ -56,3 +56,30 @@ func TestProcessUsesResolvedReportingMode(t *testing.T) {
 		})
 	}
 }
+
+func TestProcessCollectionFlags(t *testing.T) {
+	dir := t.TempDir()
+	for name, contents := range map[string]string{
+		"config.yml":      "version: 3\npipeline:\n  stages: []\nreporting:\n  enabled: false\nglossary_collections:\n  party: {entries: []}\n",
+		"transcript.json": `[{"speaker":"Alice","start":0,"end":1,"text":"hello"}]`,
+	} {
+		if err := os.WriteFile(filepath.Join(dir, name), []byte(contents), 0644); err != nil {
+			t.Fatal(err)
+		}
+	}
+	app, err := buildApplication()
+	if err != nil {
+		t.Fatal(err)
+	}
+	for _, tc := range []struct {
+		value  string
+		status int
+	}{{`["party"]`, 0}, {`[]`, 0}, {`null`, 0}, {`[1]`, 2}, {`"party"`, 2}} {
+		var out, stderr bytes.Buffer
+		args := []string{"process", "--config", filepath.Join(dir, "config.yml"), "--glossaries", tc.value, filepath.Join(dir, "transcript.json")}
+		outcome := app.Execute(context.Background(), args, commandkit.Streams{Out: &out, Err: &stderr})
+		if outcome.Status != tc.status {
+			t.Fatalf("value=%q status=%d out=%q err=%q", tc.value, outcome.Status, out.String(), stderr.String())
+		}
+	}
+}

internal/cli/process_inputs_test.go

@@ -76,3 +76,33 @@ func TestPrepareProcessInitializesReportsAndToleratesDebugLoss(t *testing.T) {
 		t.Fatalf("output published early: %v", err)
 	}
 }
+
+func TestCollectionFlagJSON(t *testing.T) {
+	for _, tc := range []struct {
+		raw   string
+		valid bool
+		null  bool
+		size  int
+	}{
+		{`["party","spells"]`, true, false, 2}, {`[]`, true, false, 0}, {`null`, true, true, 0},
+		{`"party"`, false, false, 0}, {`[1]`, false, false, 0}, {`["x"] true`, false, false, 0},
+	} {
+		result, err := parseCollectionNames(tc.raw)
+		if (err == nil) != tc.valid {
+			t.Fatalf("%s error=%v", tc.raw, err)
+		}
+		if !tc.valid {
+			continue
+		}
+		if tc.null {
+			if result != nil {
+				t.Fatalf("null=%v", result)
+			}
+			continue
+		}
+		list, ok := result.([]any)
+		if !ok || len(list) != tc.size {
+			t.Fatalf("%s result=%v", tc.raw, result)
+		}
+	}
+}