Add GLM SDK Implementation

- Introduced a new GLM SDK for interacting with the GLM (Zhipu AI) API, designed for chat completions and text generation tasks.
- Implemented core components including Client, Service, and SDK layers following Clean Architecture principles.
- Added configuration management, error handling, and structured logging for improved usability and maintainability.
- Included comprehensive documentation and usage examples to facilitate integration and understanding of the SDK's functionality.
- Enhanced the API with features such as chat completion, streaming responses, and model management, ensuring robust interaction with the GLM service.

pkg/glm/errors.go

@@ -0,0 +1,94 @@
+package glm
+
+import (
+	"fmt"
+)
+
+// Error represents a GLM SDK error
+type Error struct {
+	Code    string
+	Message string
+	Cause   error
+}
+
+// Error implements the error interface
+func (e *Error) Error() string {
+	if e.Cause != nil {
+		return fmt.Sprintf("glm: %s: %s: %v", e.Code, e.Message, e.Cause)
+	}
+	return fmt.Sprintf("glm: %s: %s", e.Code, e.Message)
+}
+
+// Unwrap returns the underlying cause of the error
+func (e *Error) Unwrap() error {
+	return e.Cause
+}
+
+// NewError creates a new GLM error
+func NewError(code, message string, cause error) *Error {
+	return &Error{
+		Code:    code,
+		Message: message,
+		Cause:   cause,
+	}
+}
+
+// Error codes
+const (
+	ErrCodeInvalidConfig       = "INVALID_CONFIG"
+	ErrCodeHTTPError          = "HTTP_ERROR"
+	ErrCodeInvalidRequest     = "INVALID_REQUEST"
+	ErrCodeAuthentication     = "AUTHENTICATION_ERROR"
+	ErrCodeRateLimit         = "RATE_LIMIT_ERROR"
+	ErrCodeServerError       = "SERVER_ERROR"
+	ErrCodeInvalidResponse   = "INVALID_RESPONSE"
+	ErrCodeChatCompletion    = "CHAT_COMPLETION_ERROR"
+	ErrCodeModelNotFound     = "MODEL_NOT_FOUND"
+	ErrCodeTokenLimit        = "TOKEN_LIMIT_ERROR"
+	ErrCodeStreamError       = "STREAM_ERROR"
+)
+
+// Error constructors
+func ErrInvalidConfig(message string) *Error {
+	return NewError(ErrCodeInvalidConfig, message, nil)
+}
+
+func ErrHTTPError(statusCode int, message string) *Error {
+	return NewError(ErrCodeHTTPError, fmt.Sprintf("HTTP %d: %s", statusCode, message), nil)
+}
+
+func ErrInvalidRequest(message string) *Error {
+	return NewError(ErrCodeInvalidRequest, message, nil)
+}
+
+func ErrAuthentication(message string) *Error {
+	return NewError(ErrCodeAuthentication, message, nil)
+}
+
+func ErrRateLimit(message string) *Error {
+	return NewError(ErrCodeRateLimit, message, nil)
+}
+
+func ErrServerError(message string) *Error {
+	return NewError(ErrCodeServerError, message, nil)
+}
+
+func ErrInvalidResponse(message string) *Error {
+	return NewError(ErrCodeInvalidResponse, message, nil)
+}
+
+func ErrChatCompletion(message string, cause error) *Error {
+	return NewError(ErrCodeChatCompletion, message, cause)
+}
+
+func ErrModelNotFound(model string) *Error {
+	return NewError(ErrCodeModelNotFound, fmt.Sprintf("model '%s' not found", model), nil)
+}
+
+func ErrTokenLimit(message string) *Error {
+	return NewError(ErrCodeTokenLimit, message, nil)
+}
+
+func ErrStreamError(message string, cause error) *Error {
+	return NewError(ErrCodeStreamError, message, cause)
+}