@brief Normalize messages by merging consecutive user messages (like Ollama does) @param messages the original messages @return normalized messages with consecutive user messages merged
| 24 | ///@param messages the original messages |
| 25 | ///@return normalized messages with consecutive user messages merged |
| 26 | static json normalize_messages(json messages) { |
| 27 | if (messages.empty()) return messages; |
| 28 | |
| 29 | json normalized = nlohmann::ordered_json::array(); |
| 30 | |
| 31 | for (size_t i = 0; i < messages.size(); i++) { |
| 32 | auto current_msg = messages[i]; |
| 33 | std::string role = current_msg.value("role", ""); |
| 34 | |
| 35 | if (role == "user") { |
| 36 | json merged_content_array = json::array(); |
| 37 | |
| 38 | // Merge all consecutive user messages into array format |
| 39 | while (i < messages.size() && messages[i].value("role", "") == "user") { |
| 40 | if (messages[i].contains("content")) { |
| 41 | if (messages[i]["content"].is_array()) { |
| 42 | for (auto& item : messages[i]["content"]) { |
| 43 | merged_content_array.push_back(item); |
| 44 | } |
| 45 | } |
| 46 | else if (messages[i]["content"].is_string()) { |
| 47 | std::string text = messages[i]["content"].get<std::string>(); |
| 48 | if (!text.empty()) { |
| 49 | nlohmann::ordered_json text_item; |
| 50 | text_item["type"] = "text"; |
| 51 | text_item["text"] = text; |
| 52 | merged_content_array.push_back(text_item); |
| 53 | } |
| 54 | } |
| 55 | } |
| 56 | if (i + 1 < messages.size() && messages[i + 1].value("role", "") == "user") i++; |
| 57 | else break; |
| 58 | } |
| 59 | |
| 60 | current_msg["content"] = merged_content_array; |
| 61 | } |
| 62 | else if (role == "system") { |
| 63 | json merged_content_array = json::array(); |
| 64 | |
| 65 | // Merge all consecutive system messages into array format |
| 66 | while (i < messages.size() && messages[i].value("role", "") == "system") { |
| 67 | if (messages[i].contains("content")) { |
| 68 | if (messages[i]["content"].is_array()) { |
| 69 | for (auto& item : messages[i]["content"]) { |
| 70 | merged_content_array.push_back(item); |
| 71 | } |
| 72 | } |
| 73 | else if (messages[i]["content"].is_string()) { |
| 74 | std::string text = messages[i]["content"].get<std::string>(); |
| 75 | if (!text.empty()) { |
| 76 | json text_item; |
| 77 | text_item["type"] = "text"; |
| 78 | text_item["text"] = text; |
| 79 | merged_content_array.push_back(text_item); |
| 80 | } |
| 81 | } |
| 82 | } |
| 83 | if (i + 1 < messages.size() && messages[i + 1].value("role", "") == "system") i++; |