Align Go ingestion boundaries with Python (#16647)

Moves doc_id blob resolution into Parser, tightens chunker/tokenizer to
Python output_format semantics, updates extractor list handling, and
fixes real-template integration tests.

internal/parser/parser/pdf_parser_nocgo_test.go

@@ -0,0 +1,31 @@
+//go:build !cgo
+
+package parser
+
+import (
+	"errors"
+	"testing"
+)
+
+func TestPDFParser_ParseWithResult_NoCGO(t *testing.T) {
+	pdf := NewPDFParser()
+
+	empty := pdf.ParseWithResult("empty.pdf", nil)
+	if empty.Err != nil {
+		t.Fatalf("empty input: want nil err, got %v", empty.Err)
+	}
+	if empty.OutputFormat != "json" {
+		t.Fatalf("empty input OutputFormat = %q, want json", empty.OutputFormat)
+	}
+	if len(empty.JSON) != 1 {
+		t.Fatalf("empty input JSON len = %d, want 1", len(empty.JSON))
+	}
+
+	res := pdf.ParseWithResult("a.pdf", []byte("%PDF-1.4"))
+	if res.Err == nil {
+		t.Fatal("want ErrPDFEngineUnavailable, got nil")
+	}
+	if !errors.Is(res.Err, ErrPDFEngineUnavailable) {
+		t.Fatalf("err = %v, want wraps ErrPDFEngineUnavailable", res.Err)
+	}
+}