ssc_codegen 0.36.1__tar.gz → 0.36.3__tar.gz

{ssc_codegen-0.36.1 → ssc_codegen-0.36.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: ssc_codegen
-Version: 0.36.1
+Version: 0.36.3
 Summary: Python-dsl code converter to html parser for web scraping 
 Project-URL: Documentation, https://github.com/vypivshiy/selector_schema_codegen#readme
 Project-URL: Issues, https://github.com/vypivshiy/selector_schema_codegen/issues

{ssc_codegen-0.36.1 → ssc_codegen-0.36.3}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "ssc_codegen"
-version = "0.36.1"
+version = "0.36.3"
 description = "Python-dsl code converter to html parser for web scraping "
 readme = "README.md"
 requires-python = ">=3.10"

{ssc_codegen-0.36.1 → ssc_codegen-0.36.3}/ssc_codegen/targets/golang/rest.py

@@ -20,7 +20,9 @@ from ssc_codegen.ast import (
     PlaceholderTemplate,
     ResultAliasDef,
     ResultVariantDef,
+    Struct,
     StructBase,
+    StructType as ST,
 )
 from ssc_codegen.ast.struct import RequestHttp
 from ssc_codegen.naming import to_camel_case, to_pascal_case, to_snake_case
@@ -338,7 +340,13 @@ def emit_method_fetch(
             )
 
     # Construct parser from body.
-    lines.append(f"{i2}return New{name}(string(body))")
+    # (raw)struct constructor returns a single value (*Name), wrapper
+    # signature is (*Name, error) → must add `, nil`. HTML constructor
+    # already returns (*Name, error) so naked tuple pass-through works.
+    if isinstance(struct, Struct) and struct.type == ST.RAW:
+        lines.append(f"{i2}return New{name}(string(body)), nil")
+    else:
+        lines.append(f"{i2}return New{name}(string(body))")
     lines.append(f"{i1}}}")
     lines.append("")
     return lines

{ssc_codegen-0.36.1 → ssc_codegen-0.36.3}/ssc_codegen/targets/golang/visitor.py

@@ -138,7 +138,6 @@ from ssc_codegen.naming import (
 from ssc_codegen.traversal.utils import (
     find_predicate_container,
     module_has_html_struct,
-    module_has_rest,
     module_uses_http,
 )
 from ssc_codegen.generation.builder import ModuleBuilder
@@ -231,15 +230,17 @@ def _gofmt(source: str) -> str:
     if not gofmt:
         return source
     try:
+        # Pass UTF-8 explicitly — Windows text-mode defaults to the OEM
+        # codepage (cp1251/cp437), corrupting non-ASCII bytes in @doc
+        # comments and source strings.
         proc = subprocess.run(
             [gofmt],
-            input=source,
+            input=source.encode("utf-8"),
             capture_output=True,
-            text=True,
             timeout=10,
         )
         if proc.returncode == 0 and proc.stdout:
-            return proc.stdout
+            return proc.stdout.decode("utf-8")
     except (subprocess.SubprocessError, OSError):
         pass
     return source
@@ -321,7 +322,7 @@ class GoVisitor(BaseWalker):
         for imp in self._builder.std_imports:
             if imp not in self._all_std_imports:
                 self._all_std_imports.append(imp)
-        if module_has_rest(module_ast):
+        if module_uses_http(module_ast):
             self._has_rest = True
         out: dict[str, str] = {"": _gofmt("\n".join(lines))}
         for fname, provider in self._file_providers.items():
@@ -1358,6 +1359,14 @@ class GoVisitor(BaseWalker):
         if at and at.base == VT.STRING and not at.is_array:
             idx = f"len({ctx.prv})-{abs(i)}" if i < 0 else str(i)
             return [f"{ctx.indent}{ctx.nxt} := string({ctx.prv}[{idx}])"]
+        if at and at.base == VT.DOCUMENT:
+            # *goquery.Selection is not indexable in Go — use .Eq().
+            # Negative index → Length() + i (i is negative, e.g. -1 → last).
+            if i < 0:
+                idx = f"{ctx.prv}.Length() + {i}"
+            else:
+                idx = str(i)
+            return [f"{ctx.indent}{ctx.nxt} := {ctx.prv}.Eq({idx})"]
         if i < 0:
             return [f"{ctx.indent}{ctx.nxt} := {ctx.prv}[len({ctx.prv})+{i}]"]
         return [f"{ctx.indent}{ctx.nxt} := {ctx.prv}[{i}]"]