fix(translator): address PR feedback for tool name sanitization

- Pre-compile sanitization regex for better performance.
- Optimize SanitizeFunctionName for conciseness and correctness.
- Handle 64-char edge cases by truncating before prepending underscore.
- Fix bug in Antigravity translator (incorrect join index).
- Refactor Gemini translators to avoid redundant sanitization calls.
- Add comprehensive unit tests including 64-char edge cases.

Co-authored-by: factory-droid[bot] <138933559+factory-droid[bot]@users.noreply.github.com>
This commit is contained in:
Saboor Hassan
2025-12-31 01:54:41 +05:00
parent f4d4249ba5
commit d241359153
5 changed files with 22 additions and 17 deletions

View File

@@ -15,28 +15,29 @@ import (
log "github.com/sirupsen/logrus"
)
var functionNameSanitizer = regexp.MustCompile(`[^a-zA-Z0-9_.:-]`)
// SanitizeFunctionName ensures a function name matches the requirements for Gemini/Vertex AI.
// It replaces invalid characters with underscores, ensures it starts with a letter or underscore,
// and truncates it to 64 characters if necessary.
// Regex Rule: [^a-zA-Z0-9_.:-] replaced with _.
func SanitizeFunctionName(name string) string {
if name == "" {
return name
return ""
}
// Replace invalid characters with underscore
re := regexp.MustCompile(`[^a-zA-Z0-9_.:-]`)
sanitized := re.ReplaceAllString(name, "_")
sanitized := functionNameSanitizer.ReplaceAllString(name, "_")
// Ensure it starts with a letter or underscore
if len(sanitized) > 0 {
first := sanitized[0]
if !((first >= 'a' && first <= 'z') || (first >= 'A' && first <= 'Z') || first == '_') {
// If it starts with an allowed character but not allowed at the beginning,
// we must prepend an underscore.
sanitized = "_" + sanitized
first := sanitized[0]
if !((first >= 'a' && first <= 'z') || (first >= 'A' && first <= 'Z') || first == '_') {
// If it starts with an allowed character but not allowed at the beginning,
// we must prepend an underscore.
// To stay within the 64-character limit while prepending, we may need to truncate first.
if len(sanitized) >= 64 {
sanitized = sanitized[:63]
}
} else {
sanitized = "_"
sanitized = "_" + sanitized
}
// Truncate to 64 characters