vectorgov-cli 0.2.6__tar.gz → 0.2.8__tar.gz

{vectorgov_cli-0.2.6 → vectorgov_cli-0.2.8}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: vectorgov-cli
-Version: 0.2.6
+Version: 0.2.8
 Summary: CLI para a API VectorGov - Busca semântica em legislação brasileira
 Project-URL: Homepage, https://vectorgov.io
 Project-URL: Documentation, https://vectorgov.io/documentacao

{vectorgov_cli-0.2.6 → vectorgov_cli-0.2.8}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "vectorgov-cli"
-version = "0.2.6"
+version = "0.2.8"
 description = "CLI para a API VectorGov - Busca semântica em legislação brasileira"
 readme = "README.md"
 license = "MIT"

{vectorgov_cli-0.2.6 → vectorgov_cli-0.2.8}/src/vectorgov/cli/__init__.py

@@ -7,7 +7,7 @@ Uso:
     vectorgov auth login
 """
 
-__version__ = "0.2.6"
+__version__ = "0.2.8"
 
 from .main import app
 

{vectorgov_cli-0.2.6 → vectorgov_cli-0.2.8}/src/vectorgov/cli/commands/lookup.py

@@ -130,6 +130,72 @@ def lookup(  # noqa: C901
 
         status = getattr(result, "status", None)
 
+        # Backend faz auto-split quando a query contem multiplas referencias
+        # separadas por virgula ou "e" (ex: "inc. I do Art. 4, inc. II do Art. 4").
+        # Nesse caso retorna status=batch com a lista em `results`.
+        if status == "batch":
+            batch_results = getattr(result, "results", None) or []
+            if raw or effective_output == "raw":
+                data = {
+                    "batch": True,
+                    "total": len(batch_results),
+                    "results": [],
+                }
+                for r in batch_results:
+                    raw_r = getattr(r, "_raw_response", None) or {}
+                    data["results"].append({
+                        "reference": getattr(r, "query", ""),
+                        "status": getattr(r, "status", "unknown"),
+                        "text": raw_r.get("text") or getattr(r, "stitched_text", "") or "",
+                        "evidence_url": absolute_url(
+                            getattr(r, "evidence_url", None) or raw_r.get("evidence_url")
+                        ),
+                        "document_url": absolute_url(
+                            getattr(r, "document_url", None) or raw_r.get("document_url")
+                        ),
+                        "document_id": raw_r.get("document_id", ""),
+                        "device_type": raw_r.get("device_type", ""),
+                        "breadcrumb": raw_r.get("breadcrumb", ""),
+                    })
+                print(json.dumps(data, ensure_ascii=False, indent=2))
+                return
+
+            # Text / json / llm output
+            if effective_output == "llm":
+                for i, r in enumerate(batch_results):
+                    raw_r = getattr(r, "_raw_response", None) or {}
+                    r_status = getattr(r, "status", "unknown")
+                    r_text = raw_r.get("text") or getattr(r, "stitched_text", "") or ""
+                    r_ref = getattr(r, "query", f"ref {i+1}")
+                    r_ev = absolute_url(getattr(r, "evidence_url", None) or raw_r.get("evidence_url"))
+                    r_doc = absolute_url(getattr(r, "document_url", None) or raw_r.get("document_url"))
+                    print(f"[{i+1}/{len(batch_results)}] {r_ref} ({r_status})")
+                    if r_text:
+                        print(r_text)
+                    if r_ev:
+                        print(f"EVIDENCE: {r_ev}")
+                    if r_doc:
+                        print(f"PDF: {r_doc}")
+                    if i < len(batch_results) - 1:
+                        print("---")
+            else:
+                console.print(f"[bold]Batch:[/bold] {len(batch_results)} referencias encontradas")
+                for i, r in enumerate(batch_results):
+                    raw_r = getattr(r, "_raw_response", None) or {}
+                    r_status = getattr(r, "status", "unknown")
+                    r_text = raw_r.get("text") or getattr(r, "stitched_text", "") or ""
+                    r_ref = getattr(r, "query", f"ref {i+1}")
+                    r_breadcrumb = raw_r.get("breadcrumb", "")
+                    title = f"[{i+1}] {r_ref} [{r_status}]"
+                    if r_breadcrumb:
+                        title += f"\n  {r_breadcrumb}"
+                    if r_text:
+                        console.print(Panel(r_text, title=title, border_style="cyan"))
+                    else:
+                        console.print(f"\n[yellow]{title}[/yellow]")
+                        console.print(f"  Nenhum texto retornado.")
+            return
+
         # Fallback: SDK v0.16 falha ao parsear `match` quando o backend aplica
         # _strip_lookup_internals (remove node_id/span_id por segurança). O
         # backend envia text/document_id/device_type no root do response mas
@@ -279,22 +345,44 @@ def lookup(  # noqa: C901
             print(render_llm_output(hits, query, metadata))
             return
 
+        # hits pode conter dict (fallback _build_match_from_raw) OU objeto
+        # (SDK Hit). Helper para ler campos de ambos uniformemente — usar
+        # getattr direto em dict não lê chaves, então str(h) acabava
+        # serializando o dict inteiro dentro do campo `text` do output JSON.
+        def _get(obj, key, default=""):
+            if obj is None:
+                return default
+            if isinstance(obj, dict):
+                return obj.get(key, default)
+            return getattr(obj, key, default)
+
         if output == "json":
-            data = [
-                {
-                    "source": getattr(h, "source", ""),
-                    "text": getattr(h, "text", str(h)),
-                    "span_id": getattr(h, "span_id", ""),
-                }
-                for h in hits
-            ]
+            data = {
+                "status": status,
+                "query": query,
+                "evidence_url": evidence_url,
+                "document_url": document_url,
+                "total": len(hits),
+                "hits": [
+                    {
+                        "source": _get(h, "source") or _get(h, "document_id"),
+                        "text": _get(h, "text"),
+                        "span_id": _get(h, "span_id"),
+                        "device_type": _get(h, "device_type"),
+                        "document_id": _get(h, "document_id"),
+                        "article_number": _get(h, "article_number"),
+                        "breadcrumb": _get(h, "breadcrumb"),
+                    }
+                    for h in hits
+                ],
+            }
             console.print_json(json.dumps(data, ensure_ascii=False))
         else:
             for h in hits:
-                source = getattr(h, "source", getattr(h, "document_id", ""))
-                span = getattr(h, "span_id", "")
+                source = _get(h, "source") or _get(h, "document_id")
+                span = _get(h, "span_id")
                 title = f"{source} — {span}" if span else source
-                text = getattr(h, "text", str(h))
+                text = _get(h, "text") or str(h)
                 console.print(Panel(text, title=title, border_style="cyan"))
 
             # Links de evidência (extraídos do result, não dos hits — o backend