[BUGFIX] handle package name more gracefully

[chensjlv]

PR Summary: Handle Package Name More Gracefully

Overview

This commit fixes a critical bug in field resolution where long package names caused incorrect message name extraction.

Changes Made

File: protobuf_to_pydantic/gen_model.py

• Modified _get_field_info_dict_by_full_name() to accept package_name parameter
• Added logic to strip package prefix from full names during field resolution
• Updated related method signatures to pass package name consistently
• Added package name extraction from descriptor file

Example Scenario

Consider a protobuf file with:

syntax = "proto3";
package analytics.dataplatform.user;

message UserProfile {
  string user_id = 1;
  int32 age = 2;
}

Why Previous Implementation Was Buggy

The old code would split analytics.dataplatform.user.UserProfile.user_id into ["analytics", "dataplatform", "user", "UserProfile", "user_id"] and incorrectly extract dataplatform as the message name (line 211: message_name, *key_list = split_full_name[1:]). This caused field validation rules to be lost since the lookup would search for dataplatform.user_id instead of UserProfile.user_id.

Benefits

• Accurate Field Resolution: Properly handles long package names in field lookups
• Preserved Validation Rules: Ensures field validation metadata is correctly applied
• Consistent Behavior: Uniform package name handling across all message types

[miloqinCedar]

this would solve my case too
very nice to have

[so1n]

Thanks for your PR. It would be great if you could add test cases. @chensjlv