Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/AngelMoralesChazari/TautoTeacher-2.0/llms.txt

Use this file to discover all available pages before exploring further.

LogicScriptEngine (package tautoteacher2.logicscript) is the central orchestrator that converts a Spanish natural-language sentence into a propositional formula string. It wires together four pipeline stages — NormalizadorTexto, NaturalLexer, SemanticMapper, and EmitidorFormula — and manages the loading of core.lgs so that callers never have to deal with classpath resources directly. The engine is never used directly from application code; instead, all callers go through LogicScriptService, the public facade that wraps a single engine instance.

​
LogicScriptEngine.traducir(String textoOriginal)

The traducir method is the single entry point for the entire pipeline. It always returns a LogicScriptResult — never throws.
1

Load core.lgs

At construction time the engine already called LgsCargador.cargarConDiagnostico("logicscript/core.lgs") and stored the result. At the start of traducir, it inspects that result. If bloqueaTraduccion() is true (parse error or unrecoverable I/O error), the method records the load error in pasosDeAnalisis and immediately returns LogicScriptResult.error(...). If the status is RECURSO_NO_ENCONTRADO, it adds a warning step and continues using the Java-embedded fallback lemmas and patterns.
2

Normalize text

NormalizadorTexto.normalizar(textoOriginal) is called to lowercase the input, strip diacritics where applicable, and produce a clean string. If the result is empty, an error result is returned. The normalized string is added to pasosDeAnalisis.
3

Tokenize

NaturalLexer.tokenizar(texto) converts the normalized string into a List<TokenNatural>. Each token is either a keyword (SI, ENTONCES, Y, O, SIEMPRE_QUE, CUANDO, EN_CASO_DE_QUE, SOLO_SI, SI_Y_SOLO_SI, A_MENOS_QUE) or a LITERAL carrying the span text. The full token list is added to pasosDeAnalisis.
4

Segment compound sentences

The normalized text is split into independent clauses on the separators , si, , siempre que, , cuando, and , en caso de que. Each segment becomes its own translation unit. If no segments are produced, an error result is returned. When more than one segment exists, a composition note is added to pasosDeAnalisis.
5

Translate each segment

For every segment, SemanticMapper.mapearBloque(bloque, tokensBloque, pasosDeAnalisis) is called. The mapper tries to match the token list against the patterns declared in core.lgs (in declaration order, first match wins). If no pattern matches, the segment falls back to a bare atom via BaseConocimiento.canonicalizarFragmento. A null result from the mapper causes the whole translation to fail.
6

Join segments with conjunction

If there are multiple translated segments, they are combined left-to-right using AndExpr nodes so the result is a single LogicExpr tree representing (seg1 ∧ seg2 ∧ …).
7

Emit formula string

EmitidorFormula.emitir(expresionFinal, registro, pasosDeAnalisis) walks the LogicExpr tree, assigns short proposition symbols (p, q, r, …) via RegistroProposiciones, and serializes the result using Unicode operators (→, ∧, ∨, ↔, ¬).
8

Return result

LogicScriptResult.exito(formula, mensaje, pasosDeAnalisis, registro.mapaParaResultado()) is returned with the formula string, a success message, the full analysis trace, and the proposition map.

​
LogicScriptResult fields

​
exito
boolean
Whether the translation pipeline completed successfully. Check this before reading formula.
​
formula
String
The emitted propositional formula using Unicode symbols (e.g. (llover → llevar)). Empty string when exito is false.
​
mensaje
String
A human-readable status message — "Traducción LogicScript completada." on success, or a description of the error on failure.
​
pasosDeAnalisis
List<String>
An unmodifiable ordered list of trace entries recorded at each pipeline stage: the normalized input, the raw token list, the block being processed, matched pattern names, fallback decisions, and any load warnings. Very useful for debugging.
​
proposiciones
Map<String, String>
An unmodifiable map from short symbol to canonical fragment, e.g. {p=llover, q=llevar}. Empty when exito is false.

​
LogicScriptService

LogicScriptService is a thin, stateless-from-the-caller’s-perspective facade that creates one LogicScriptEngine instance at construction time and exposes a single method: 
public LogicScriptResult traducir(String textoNatural)
Both the GUI (MainWindow) and the CLI (LogicScriptCli) instantiate LogicScriptService and call traducir. This ensures that the same pipeline — including the same loaded core.lgs data — is used in every context.

​
Java usage example

LogicScriptService service = new LogicScriptService();
LogicScriptResult result = service.traducir("si llueve entonces llevo paraguas");
if (result.isExito()) {
    System.out.println(result.getFormula());       // (llover → llevar)
    System.out.println(result.getProposiciones()); // {p=llover, q=llevar}
}

​
EstadoCargaLgs enum values

EstadoCargaLgs is returned inside ResultadoCargaLgs by LgsCargador.cargarConDiagnostico. The engine inspects it at startup to decide how to proceed.
ValueMeaningEngine behaviour
EXITOFile read and parsed without errors.All declarative rules are active.
RECURSO_NO_ENCONTRADOThe .lgs path is not on the classpath.Warning logged in pasosDeAnalisis; Java-embedded lemmas and patterns used as fallback. Translation proceeds.
ERROR_SINTAXISA line in the file could not be parsed (message includes line number).bloqueaTraduccion() returns true; translation is blocked and an error result is returned immediately.
ERROR_LECTURAAn I/O failure unrelated to parsing occurred.Same blocking behaviour as ERROR_SINTAXIS.
The pasosDeAnalisis list is your first debugging tool. It captures every decision the pipeline makes — which pattern was matched, when a fallback atom was used, whether the .lgs file was found — so you can trace exactly why a sentence produced an unexpected formula. Print it to the console or display it in the GUI’s detail panel whenever a translation looks wrong.

Build docs developers (and LLMs) love

Get started for freeTalk to us