fix(osgen): treat ISM/KNN/LTR/ML/PPL/SM/UBI/WLM as acronyms

Go identifier convention requires acronyms to be all-uppercase (the
generator already does this for API, HTTP, JSON, URL, etc.). The
OpenSearch plugin-namespace acronyms were missing from the table, so the
generated surface emitted IsmPolicy, KnnStats, MlModel, SmPolicy, and
similar mis-cased identifiers.

Add the missing entries to the acronyms map so titleSegment expands them:

  ISM  Index State Management
  KNN  k-Nearest Neighbors
  LTR  Learning to Rank
  ML   Machine Learning
  PPL  Piped Processing Language
  SM   Snapshot Management
  UBI  User Behavior Insights
  WLM  Workload Management

Matching is whole-segment only (segments are split on '.' and '_' before
lookup), so words like "smile" are unaffected; a regression test pins
this. The rename takes effect on the next code generation, which rewrites
the affected *_gen.go identifiers (e.g. IsmPolicy -> ISMPolicy) in one
pass.

Closes opensearch-project#863

Signed-off-by: Sean Chittenden <sean.chittenden@crowdstrike.com>

Author: sean-

cmd/osgen/naming.go

@@ -10,6 +10,7 @@ import (
 	"fmt"
 	"go/token"
 	"go/types"
+	"sort"
 	"strings"
 )
 
@@ -45,12 +46,67 @@ var idiomaticAbbreviations = []struct {
 // s, matching each pattern at PascalCase boundaries (followed by
 // uppercase, or end-of-string for entries with tailUpperOnly=false).
 func applyIdiomaticAbbreviations(s string) string {
+	// Normalize embedded acronyms first. A single spec token like
+	// "IsmTemplate" pascal-cases to "IsmTemplate" (titleSegment only expands
+	// whole, separately-delimited segments), leaving the acronym in mixed
+	// case. Canonicalizing "Ism" -> "ISM" here -- at PascalCase boundaries --
+	// keeps acronym casing consistent wherever it appears in an identifier
+	// (prefix or local part), which also lets deStutterPrefix match.
+	for _, a := range acronymBoundaryReplacements() {
+		s = replaceAtPascalBoundary(s, a.from, a.to, a.tailUpperOnly)
+	}
 	for _, a := range idiomaticAbbreviations {
 		s = replaceAtPascalBoundary(s, a.from, a.to, a.tailUpperOnly)
 	}
 	return s
 }
 
+// acronymBoundaryReplacements derives, from the acronyms table, the
+// substitutions that canonicalize a title-cased embedded acronym (e.g. "Ism",
+// "Knn") to its idiomatic all-caps form (e.g. "ISM", "KNN"). Two-letter and
+// longer acronyms qualify; single-letter "acronyms" would over-match, so they
+// are skipped. Each is applied at PascalCase boundaries (followed by an
+// uppercase letter or end-of-string), so "Ismael"-style words are untouched.
+//
+//nolint:gochecknoglobals // derived once from the acronyms table
+var acronymBoundaryReplacementsCache []struct {
+	from, to      string
+	tailUpperOnly bool
+}
+
+func acronymBoundaryReplacements() []struct {
+	from, to      string
+	tailUpperOnly bool
+} {
+	if acronymBoundaryReplacementsCache != nil {
+		return acronymBoundaryReplacementsCache
+	}
+	for lower, upper := range acronyms {
+		if len(lower) < 2 {
+			continue
+		}
+		title := strings.ToUpper(lower[:1]) + lower[1:]
+		if title == upper {
+			continue // already idiomatic (e.g. all-lowercase has no title form to fix)
+		}
+		acronymBoundaryReplacementsCache = append(acronymBoundaryReplacementsCache, struct {
+			from, to      string
+			tailUpperOnly bool
+		}{from: title, to: upper})
+	}
+	// Deterministic order: map iteration is random, and codegen output must be
+	// stable across runs. Longest-from first so multi-token acronyms can't be
+	// partially shadowed by a shorter one.
+	sort.Slice(acronymBoundaryReplacementsCache, func(i, j int) bool {
+		a, b := acronymBoundaryReplacementsCache[i].from, acronymBoundaryReplacementsCache[j].from
+		if len(a) != len(b) {
+			return len(a) > len(b)
+		}
+		return a < b
+	})
+	return acronymBoundaryReplacementsCache
+}
+
 // replaceAtPascalBoundary replaces every occurrence of old in s with
 // next, but only when old is followed by an uppercase letter -- or
 // end-of-string if tailUpperOnly is false. Lowercase-suffix variants
@@ -89,16 +145,24 @@ var acronyms = map[string]string{
 	"https": "HTTPS",
 	"id":    "ID",
 	"ip":    "IP",
+	"ism":   "ISM", // Index State Management
 	"json":  "JSON",
-	"pit":   "PIT",
+	"knn":   "KNN", // k-Nearest Neighbors
+	"ltr":   "LTR", // Learning to Rank
+	"ml":    "ML",  // Machine Learning
+	"pit":   "PIT", // Point In Time
+	"ppl":   "PPL", // Piped Processing Language
+	"sm":    "SM",  // Snapshot Management
 	"sql":   "SQL",
 	"ssl":   "SSL",
 	"tcp":   "TCP",
 	"tls":   "TLS",
 	"ttl":   "TTL",
+	"ubi":   "UBI", // User Behavior Insights
 	"uri":   "URI",
 	"url":   "URL",
 	"uuid":  "UUID",
+	"wlm":   "WLM", // Workload Management
 	"xml":   "XML",
 }
 

cmd/osgen/naming_test.go

@@ -33,6 +33,15 @@ func TestTitleSegment(t *testing.T) {
 		{name: "ssl acronym", input: "ssl", want: "SSL"},
 		{name: "api acronym", input: "api", want: "API"},
 		{name: "json acronym", input: "json", want: "JSON"},
+		{name: "ism acronym", input: "ism", want: "ISM"},
+		{name: "knn acronym", input: "knn", want: "KNN"},
+		{name: "ltr acronym", input: "ltr", want: "LTR"},
+		{name: "ml acronym", input: "ml", want: "ML"},
+		{name: "ppl acronym", input: "ppl", want: "PPL"},
+		{name: "sm acronym", input: "sm", want: "SM"},
+		{name: "ubi acronym", input: "ubi", want: "UBI"},
+		{name: "wlm acronym", input: "wlm", want: "WLM"},
+		{name: "whole-segment only, not substring", input: "smile", want: "Smile"},
 		{name: "empty", input: "", want: ""},
 		{name: "mixed case id", input: "ID", want: "ID"},
 		{name: "mixed case uuid", input: "UUID", want: "UUID"},
@@ -276,6 +285,12 @@ func TestSchemaTypeName(t *testing.T) {
 		{name: "group._common cluster", schemaKey: "cluster._common___ComponentTemplate", want: "ClusterComponentTemplate"},
 		{name: "acronyms", schemaKey: "security._common___SSLInfo", want: "SecuritySSLInfo"},
 		{name: "sql plugin", schemaKey: "sql._common___SQLQuery", want: "SQLQuery"},
+		{name: "ism plugin acronym", schemaKey: "ism._common___Policy", want: "ISMPolicy"},
+		{name: "knn plugin acronym", schemaKey: "knn._common___Stats", want: "KNNStats"},
+		// Embedded acronym in the local part must normalize and de-stutter:
+		// "IsmTemplate" -> "ISMTemplate", not "ISMIsmTemplate".
+		{name: "ism embedded acronym de-stutters", schemaKey: "ism._common___IsmTemplate", want: "ISMTemplate"},
+		{name: "knn embedded acronym de-stutters", schemaKey: "knn._common___KnnMethod", want: "KNNMethod"},
 		{name: "de-stutter empty result kept", schemaKey: "cluster.health___Health", want: "ClusterHealthHealth"},
 
 		// Idiomatic abbreviations: M-prefix initialisms, compound nouns,