Add native El NLG system: morphology, vocabulary, grammar, realizer

Implements a complete natural language generation stack in El: - morphology.el: English pluralization, verb conjugation (40+ irregulars), determiner agreement - vocabulary.el: inline seed lexicon (~100 entries: pronouns, nouns, verbs, adjectives, etc.) - grammar.el: CFG rules (S/NP/VP/PP), slot-map driven tree generator, s-expression renderer - realizer.el: semantic form -> English text with tense/aspect/agreement, do-support for questions - nlg.el: JSON-driven public API tying all modules together - tests/run.sh: acceptance corpus runner (6 tests, all passing)
2026-05-02 14:16:23 -05:00
commit 2a6d70b959
13 changed files with 1956 additions and 0 deletions
@@ -0,0 +1,112 @@
+#!/usr/bin/env bash
+# run.sh - build and execute the nlg/ acceptance corpus.
+#
+# Each examples/<case>.el is a self-contained El program that calls NLG
+# functions and prints a deterministic result line. Because El has no
+# import system, the runner concatenates all NLG source modules above each
+# test file before handing it to elc.
+#
+# Module load order:
+#   morphology.el  (standalone)
+#   vocabulary.el  (standalone)
+#   grammar.el     (uses slot helpers)
+#   realizer.el    (uses morphology + grammar)
+#   nlg.el         (public API)
+#   <test>.el      (calls the above)
+
+set -uo pipefail
+cd "$(dirname "$0")"
+
+EL_HOME="${EL_HOME:-$(cd ../.. && pwd)/el}"
+ELC="${ELC:-${EL_HOME}/dist/platform/elc}"
+RUNTIME_DIR="${EL_HOME}/el-compiler/runtime"
+SRC_DIR="$(cd .. && pwd)/src"
+
+if [ ! -x "${ELC}" ]; then
+    echo "elc not found at ${ELC}" >&2
+    exit 1
+fi
+
+# NLG source modules in dependency order
+NLG_MODULES=(
+    "${SRC_DIR}/morphology.el"
+    "${SRC_DIR}/vocabulary.el"
+    "${SRC_DIR}/grammar.el"
+    "${SRC_DIR}/realizer.el"
+    "${SRC_DIR}/nlg.el"
+)
+
+PASS=0
+FAIL=0
+FAILED_NAMES=()
+
+run_nlg_case() {
+    local name="$1"
+    local src="$2"
+    local expected="$3"
+
+    local combined_src
+    local out_c
+    local out_bin
+    combined_src="$(mktemp -t nlg_test.XXXXXX).el"
+    out_c="$(mktemp -t nlg_test.XXXXXX).c"
+    out_bin="$(mktemp -t nlg_test.XXXXXX)"
+
+    # Concatenate modules + test file
+    for mod in "${NLG_MODULES[@]}"; do
+        cat "${mod}" >> "${combined_src}"
+        echo "" >> "${combined_src}"
+    done
+    cat "${src}" >> "${combined_src}"
+
+    if ! "${ELC}" "${combined_src}" > "${out_c}" 2>/tmp/nlg_test.elc.err; then
+        echo "FAIL  ${name} - elc emit failed:"
+        cat /tmp/nlg_test.elc.err | sed 's/^/      /'
+        FAIL=$((FAIL+1))
+        FAILED_NAMES+=("${name}")
+        rm -f "${combined_src}" "${out_c}" "${out_bin}"
+        return
+    fi
+
+    if ! cc -O2 -I "${RUNTIME_DIR}" "${out_c}" "${RUNTIME_DIR}/el_runtime.c" \
+            -lcurl -lpthread -o "${out_bin}" 2>/tmp/nlg_test.cc.err; then
+        echo "FAIL  ${name} - cc failed:"
+        cat /tmp/nlg_test.cc.err | sed 's/^/      /'
+        FAIL=$((FAIL+1))
+        FAILED_NAMES+=("${name}")
+        rm -f "${combined_src}" "${out_c}" "${out_bin}"
+        return
+    fi
+
+    local got
+    got="$("${out_bin}" 2>&1)"
+
+    if [ "${got}" = "${expected}" ]; then
+        echo "PASS  ${name}"
+        PASS=$((PASS+1))
+    else
+        echo "FAIL  ${name}  expected: '${expected}', got: '${got}'"
+        FAIL=$((FAIL+1))
+        FAILED_NAMES+=("${name}")
+    fi
+
+    rm -f "${combined_src}" "${out_c}" "${out_bin}"
+}
+
+echo "==> Running nlg acceptance corpus"
+echo
+
+run_nlg_case "basic-sentence"    examples/basic-sentence.el    "I see the cat."
+run_nlg_case "past-tense"        examples/past-tense.el        "She ran in the park."
+run_nlg_case "question"          examples/question.el          "Do you see the dog?"
+run_nlg_case "command"           examples/command.el           "Go home."
+run_nlg_case "plural-noun"       examples/plural-noun.el       "cats|children|boxes|cities|fish|leaves"
+run_nlg_case "verb-conjugation"  examples/verb-conjugation.el  "ran|walked|is|am|sleeping"
+
+echo
+echo "${PASS} passed, ${FAIL} failed"
+if [ "${FAIL}" -gt 0 ]; then
+    echo "failed: ${FAILED_NAMES[*]}"
+    exit 1
+fi
+exit 0