{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "0a8abbfa",
   "metadata": {},
   "outputs": [],
   "source": [
    "import re\n",
    "from dataclasses import dataclass\n",
    "import pprint\n",
    "from pathlib import Path\n",
    "from typing import Any, Dict, List, Optional, Tuple\n",
    "from lark import Tree, Lark\n",
    "from src.config import settings\n",
    "from lark import Lark"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "baa779f3",
   "metadata": {},
   "source": [
    "# Functions"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "23a92e13",
   "metadata": {},
   "source": [
    "# BNF to Lark"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "93d2db25",
   "metadata": {},
   "outputs": [],
   "source": [
    "grammar = r\"\"\"\n",
    "start: program\n",
    "\n",
    "program: separator* line_or_comment (separator+ line_or_comment)* separator*\n",
    "\n",
    "?line_or_comment: simple_stmt comment?\n",
    "                | compound_stmt\n",
    "                | comment\n",
    "                | BLOCK_COMMENT\n",
    "\n",
    "?separator: EOL+\n",
    "\n",
    "comment: DOC_COMMENT | LINE_COMMENT\n",
    "\n",
    "EOL: /\\r?\\n/\n",
    "\n",
    "DOC_COMMENT.2: /\\/\\/\\/[^\\r\\n]*/\n",
    "LINE_COMMENT.1: /\\/\\/[^\\r\\n]*/\n",
    "BLOCK_COMMENT: /\\/\\*[\\s\\S]*?\\*\\//\n",
    "\n",
    "?simple_stmt: assignment\n",
    "            | return_stmt\n",
    "            | system_command\n",
    "            | io_command\n",
    "            | async_command\n",
    "            | connector_cmd\n",
    "            | db_command\n",
    "            | http_command\n",
    "            | util_command\n",
    "            | modularity_cmd\n",
    "            | call_stmt\n",
    "\n",
    "?compound_stmt: function_decl\n",
    "              | if_stmt\n",
    "              | loop_stmt\n",
    "              | try_stmt\n",
    "\n",
    "assignment: identifier \"=\" expression\n",
    "\n",
    "call_stmt: identifier \"(\" argument_list? \")\"\n",
    "         | identifier \"=\" identifier \".\" identifier \"(\" argument_list? \")\"\n",
    "         | identifier \".\" identifier \"(\" argument_list? \")\"\n",
    "\n",
    "system_command: register_cmd\n",
    "              | addvar_cmd\n",
    "\n",
    "register_cmd: \"registerEndpoint\" \"(\" stringliteral \",\" stringliteral \",\" list_display \",\" stringliteral \",\" identifier \",\" identifier \")\"\n",
    "\n",
    "addvar_cmd: \"addVar\" \"(\" addvar_arg \",\" addvar_arg \")\"\n",
    "\n",
    "addvar_arg: identifier\n",
    "          | literal\n",
    "          | \"$\" identifier\n",
    "\n",
    "identifier: IDENTIFIER\n",
    "\n",
    "system_variable: \"_status\"\n",
    "\n",
    "io_command: addparam_cmd\n",
    "          | getlistlen_cmd\n",
    "          | addresult\n",
    "          | getparamlist_cmd\n",
    "\n",
    "addparam_cmd: \"addParam\" \"(\" stringliteral \",\" identifier \")\"\n",
    "getlistlen_cmd: \"getListLen\" \"(\" identifier \",\" identifier \")\"\n",
    "getparamlist_cmd: \"getQueryParamList\" \"(\" stringliteral \",\" identifier \")\"\n",
    "addresult: \"addResult\" \"(\" identifier \")\"\n",
    "\n",
    "if_stmt: \"if\" \"(\" if_condition \")\" separator block (\"else\" \"(\" \")\" separator block)? \"end\" \"(\" \")\"\n",
    "\n",
    "if_condition: if_atom \",\" if_atom \",\" stringliteral\n",
    "            | \"None\" \",\" \"None\" \",\" stringliteral\n",
    "\n",
    "if_atom: identifier\n",
    "       | literal\n",
    "\n",
    "loop_stmt: \"startLoop\" \"(\" identifier \",\" expression \",\" expression \")\" separator block \"endLoop\" \"(\" \")\"\n",
    "\n",
    "try_stmt: \"try\" \"(\" \")\" separator block \"exception\" \"(\" identifier \")\" separator block \"end\" \"(\" \")\"\n",
    "\n",
    "block: separator* line_or_comment (separator+ line_or_comment)* separator*\n",
    "\n",
    "async_command: go_stmt\n",
    "             | gather_stmt\n",
    "\n",
    "go_stmt: identifier \"=\" \"go\" identifier \"(\" argument_list? \")\"\n",
    "gather_stmt: identifier \"=\" \"gather\" \"(\" identifier (\",\" expression)? \")\"\n",
    "\n",
    "connector_cmd: connector_instantiation\n",
    "\n",
    "connector_instantiation: identifier \"=\" \"avapConnector\" \"(\" stringliteral \")\"\n",
    "\n",
    "http_command: req_post_cmd\n",
    "            | req_get_cmd\n",
    "\n",
    "req_post_cmd: \"RequestPost\" \"(\" expression \",\" expression \",\" expression \",\" expression \",\" identifier \",\" expression \")\"\n",
    "req_get_cmd: \"RequestGet\" \"(\" expression \",\" expression \",\" expression \",\" identifier \",\" expression \")\"\n",
    "\n",
    "db_command: orm_direct\n",
    "          | orm_check\n",
    "          | orm_create\n",
    "          | orm_select\n",
    "          | orm_insert\n",
    "          | orm_update\n",
    "\n",
    "orm_direct: \"ormDirect\" \"(\" expression \",\" identifier \")\"\n",
    "orm_check: \"ormCheckTable\" \"(\" expression \",\" identifier \")\"\n",
    "orm_create: \"ormCreateTable\" \"(\" expression \",\" expression \",\" expression \",\" identifier \")\"\n",
    "\n",
    "orm_select: \"ormAccessSelect\" \"(\" orm_fields \",\" expression (\",\" expression)? \",\" identifier \")\"\n",
    "\n",
    "orm_fields: \"*\"\n",
    "          | expression\n",
    "\n",
    "orm_insert: \"ormAccessInsert\" \"(\" expression \",\" expression \",\" identifier \")\"\n",
    "orm_update: \"ormAccessUpdate\" \"(\" expression \",\" expression \",\" expression \",\" expression \",\" identifier \")\"\n",
    "\n",
    "util_command: json_list_cmd\n",
    "            | crypto_cmd\n",
    "            | regex_cmd\n",
    "            | datetime_cmd\n",
    "            | stamp_cmd\n",
    "            | string_cmd\n",
    "            | replace_cmd\n",
    "\n",
    "json_list_cmd: \"variableToList\" \"(\" expression \",\" identifier \")\"\n",
    "             | \"itemFromList\" \"(\" identifier \",\" expression \",\" identifier \")\"\n",
    "             | \"variableFromJSON\" \"(\" identifier \",\" expression \",\" identifier \")\"\n",
    "             | \"AddVariableToJSON\" \"(\" expression \",\" expression \",\" identifier \")\"\n",
    "\n",
    "crypto_cmd: \"encodeSHA256\" \"(\" identifier_or_string \",\" identifier \")\"\n",
    "          | \"encodeMD5\" \"(\" identifier_or_string \",\" identifier \")\"\n",
    "\n",
    "regex_cmd: \"getRegex\" \"(\" identifier \",\" stringliteral \",\" identifier \")\"\n",
    "\n",
    "datetime_cmd: \"getDateTime\" \"(\" stringliteral \",\" expression \",\" stringliteral \",\" identifier \")\"\n",
    "\n",
    "stamp_cmd: \"stampToDatetime\" \"(\" expression \",\" stringliteral \",\" expression \",\" identifier \")\"\n",
    "         | \"getTimeStamp\" \"(\" stringliteral \",\" stringliteral \",\" expression \",\" identifier \")\"\n",
    "\n",
    "string_cmd: \"randomString\" \"(\" expression \",\" expression \",\" identifier \")\"\n",
    "\n",
    "replace_cmd: \"replace\" \"(\" identifier_or_string \",\" stringliteral \",\" stringliteral \",\" identifier \")\"\n",
    "\n",
    "function_decl: \"function\" identifier \"(\" param_list? \")\" \"{\" separator block \"}\"\n",
    "\n",
    "param_list: identifier (\",\" identifier)*\n",
    "\n",
    "return_stmt: \"return\" \"(\" expression? \")\"\n",
    "\n",
    "modularity_cmd: include_stmt\n",
    "              | import_stmt\n",
    "\n",
    "include_stmt: \"include\" stringliteral\n",
    "import_stmt: \"import\" (\"<\" identifier \">\" | stringliteral)\n",
    "\n",
    "?expression: logical_or\n",
    "\n",
    "?logical_or: logical_and (\"or\" logical_and)*\n",
    "?logical_and: logical_not (\"and\" logical_not)*\n",
    "\n",
    "?logical_not: \"not\" logical_not\n",
    "            | comparison\n",
    "\n",
    "?comparison: arithmetic (comp_op arithmetic)*\n",
    "\n",
    "comp_op: \"==\" | \"!=\" | \"<\" | \">\" | \"<=\" | \">=\" | \"in\" | \"is\"\n",
    "\n",
    "?arithmetic: term ((\"+\" | \"-\") term)*\n",
    "?term: factor ((\"*\" | \"/\" | \"%\") factor)*\n",
    "\n",
    "?factor: (\"+\" | \"-\") factor\n",
    "       | power\n",
    "\n",
    "?power: primary (\"**\" factor)?\n",
    "\n",
    "?primary: atom postfix*\n",
    "\n",
    "postfix: \".\" identifier\n",
    "       | \"[\" expression \"]\"\n",
    "       | \"[\" expression? \":\" expression? (\":\" expression?)? \"]\"\n",
    "       | \"(\" argument_list? \")\"\n",
    "\n",
    "?atom: identifier\n",
    "     | \"$\" identifier\n",
    "     | literal\n",
    "     | \"(\" expression \")\"\n",
    "     | list_display\n",
    "     | dict_display\n",
    "\n",
    "list_display: \"[\" argument_list? \"]\"\n",
    "            | \"[\" expression \"for\" identifier \"in\" expression if_clause? \"]\"\n",
    "\n",
    "if_clause: \"if\" expression\n",
    "\n",
    "dict_display: \"{\" key_datum_list? \"}\"\n",
    "\n",
    "key_datum_list: key_datum (\",\" key_datum)*\n",
    "key_datum: expression \":\" expression\n",
    "\n",
    "argument_list: expression (\",\" expression)*\n",
    "\n",
    "number: FLOATNUMBER\n",
    "      | INTEGER\n",
    "\n",
    "literal: stringliteral\n",
    "       | number\n",
    "       | boolean\n",
    "       | \"None\"\n",
    "\n",
    "boolean: \"True\" | \"False\"\n",
    "\n",
    "INTEGER: /[0-9]+/\n",
    "FLOATNUMBER: /(?:[0-9]+\\.[0-9]*|\\.[0-9]+)/\n",
    "\n",
    "stringliteral: STRING_DOUBLE\n",
    "             | STRING_SINGLE\n",
    "\n",
    "STRING_DOUBLE: /\"([^\"\\\\]|\\\\.)*\"/\n",
    "STRING_SINGLE: /'([^'\\\\]|\\\\.)*'/\n",
    "\n",
    "identifier_or_string: identifier\n",
    "                    | stringliteral\n",
    "\n",
    "IDENTIFIER: /[A-Za-z_][A-Za-z0-9_]*/\n",
    "\n",
    "%ignore /[ \\t]+/\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "95267b2a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# STRING_DOUBLE: /\"([^\"\\\\]|\\\\[\"'\\\\ntr0])*\"/\n",
    "# STRING_SINGLE: /'([^'\\\\]|\\\\[\"'\\\\ntr0])*'/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "0cab2125",
   "metadata": {},
   "outputs": [],
   "source": [
    "code = \"\"\"\n",
    "addVar(mensaje, \"Hola mundo desde AVAP\")\n",
    "addResult(mensaje)\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "6a266b2a",
   "metadata": {},
   "outputs": [],
   "source": [
    "folder = \"/home/pseco/VsCodeProjects/assistance-engine/docs/samples/\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "522bdb3b",
   "metadata": {},
   "outputs": [],
   "source": [
    "parser = Lark(grammar, parser=\"lalr\", propagate_positions=True, start=\"program\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "d3aa8026",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Parsed 26 files successfully\n",
      "Failed to parse 7 files\n",
      "\n",
      "Files processed:\n",
      "  - if_desigualdad.avap: error\n",
      "  - manejo_error_sql_critico.avap: error\n",
      "  - captura_de_listas_multiples.avap: error\n",
      "  - expresion_compleja.avap: error\n",
      "  - validacion_in_pertenece_a_lista.avap: error\n",
      "  - try_catch_request.avap: error\n",
      "  - else_estandar.avap: error\n"
     ]
    }
   ],
   "source": [
    "# Initialize dictionary to store parsed results\n",
    "parsed_files = {}\n",
    "folder_path = Path(folder)\n",
    "\n",
    "# Parse each .avap file\n",
    "for file_path in folder_path.glob(\"*.avap\"):\n",
    "    try:\n",
    "        # Read file with encoding handling and normalize line breaks\n",
    "        file_content = file_path.read_text(encoding='utf-8')\n",
    "        # Normalize all line endings to \\n\n",
    "        file_content = file_content.replace('\\r\\n', '\\n').replace('\\r', '\\n')\n",
    "        \n",
    "        ast = parser.parse(file_content)\n",
    "        \n",
    "        # Store AST with metadata\n",
    "        parsed_files[file_path.name] = {\n",
    "            \"ast\": ast,\n",
    "            \"tree\": ast.pretty(),  # Pretty-printed BNF structure\n",
    "            \"status\": \"success\"\n",
    "        }\n",
    "    except UnicodeDecodeError as e:\n",
    "        # If UTF-8 fails, try with latin-1 (more permissive)\n",
    "        try:\n",
    "            file_content = file_path.read_text(encoding='latin-1')\n",
    "            file_content = file_content.replace('\\r\\n', '\\n').replace('\\r', '\\n')\n",
    "            ast = parser.parse(file_content)\n",
    "            parsed_files[file_path.name] = {\n",
    "                \"ast\": ast,\n",
    "                \"tree\": ast.pretty(),\n",
    "                \"status\": \"success\"\n",
    "            }\n",
    "        except Exception as e2:\n",
    "            parsed_files[file_path.name] = {\n",
    "                \"error\": f\"Encoding error: {str(e2)}\",\n",
    "                \"status\": \"error\"\n",
    "            }\n",
    "    except Exception as e:\n",
    "        parsed_files[file_path.name] = {\n",
    "            \"error\": str(e),\n",
    "            \"status\": \"error\"\n",
    "        }\n",
    "\n",
    "# Display results summary\n",
    "print(f\"Parsed {len([f for f in parsed_files.values() if f['status'] == 'success'])} files successfully\")\n",
    "print(f\"Failed to parse {len([f for f in parsed_files.values() if f['status'] == 'error'])} files\")\n",
    "print(\"\\nFiles processed:\")\n",
    "for filename, result in parsed_files.items():\n",
    "    if result[\"status\"] == \"error\":\n",
    "        print(f\"  - {filename}: {result['status']}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "35b34377",
   "metadata": {},
   "outputs": [],
   "source": [
    "parsed_files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "e9e6c0fb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Tree('program', [Token('EOL', '\\n'), Tree('system_command', [Tree('addvar_cmd', [Tree('addvar_arg', [Tree('identifier', [Token('IDENTIFIER', 'mensaje')])]), Tree('addvar_arg', [Tree('literal', [Tree('stringliteral', [Token('STRING_DOUBLE', '\"Hola mundo desde AVAP\"')])])])])]), Token('EOL', '\\n'), Tree('io_command', [Tree('addresult', [Tree('identifier', [Token('IDENTIFIER', 'mensaje')])])]), Token('EOL', '\\n')])"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "parser.parse(code)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "assistance-engine",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}