Tutorial 05: Experimental Language Ingestion

Goal: ingest SVC-style sentence data and query semantic memory.

Step 1: Create SVC entries

from grilly.experimental.language.svc_loader import load_svc_entries_from_dicts

raw_entries = [
    {
        "id": "e0",
        "text": "Vaccines prevent infectious diseases.",
        "svc": {"s": "Vaccines", "v": "prevent", "c": "infectious diseases"},
        "pos": ["NOUN", "VERB", "ADJ", "NOUN", "PUNCT"],
        "deps": ["nsubj", "ROOT", "amod", "dobj", "punct"],
        "lemmas": ["vaccine", "prevent", "infectious", "disease", "."],
        "root_verb": "prevent",
        "realm": "health",
        "source": "manual",
        "complexity": 0.4,
    }
]
entries = load_svc_entries_from_dicts(raw_entries)

Step 2: Initialize InstantLanguage

from grilly.experimental.language.system import InstantLanguage

lang = InstantLanguage(dim=1024)

Step 3: Ingest entries

result = lang.ingest_svc(entries)
print("sentences learned:", result.sentences_learned)
print("realms:", result.realm_counts)

Step 4: Query similar sentences

query = "Vaccines are effective."
similar = lang.find_similar_sentences(query, top_k=3)
print(similar)

Step 5: Integrate with controller

from grilly.experimental.cognitive.controller import CognitiveController

controller = CognitiveController(dim=1024)
controller.ingest_svc(entries)
response = controller.process("What helps prevent diseases?")
print(response)

This workflow gives you a full path from structured sentence ingestion to cognitive querying.