Fix path escaping and label collision in batch naming

Two defects that both fail quietly rather than loudly.

freecad_prepare built the FreeCAD script by hand-escaping paths into
single quotes, but the apostrophe branch of that escaping was a no-op:
in Python source "\'" is simply "'", so the replace swapped an
apostrophe for itself. A STEP file under a path like
C:\Users\Dan's Files\part.step closed the string literal early and the
generated script raised a SyntaxError, which reached the user as an
unexplained "FreeCAD conversion failed". Paths are now interpolated
with %r, which escapes quotes and backslashes together.

alpha_pair restarted at two letters after ZZ, so index 26 produced 'AA'
again -- the same label as index 0. With Products per Letter set low
enough to wrap, two products resolved to one output directory and the
second overwrote the first's renders. The label width now grows with
the wrap count, giving AA..ZZ then AAA..ZZZ. Labels AA..ZZ are
unchanged, so existing output naming is unaffected.

Author: Danyals-code

multiview_product_renderer/__init__.py

@@ -523,13 +523,16 @@ def find_layer_coll(layer_coll, name):
 
 
 def alpha_pair(idx):
+    """Repeated-letter group label: AA..ZZ, then AAA..ZZZ, then AAAA..
+
+    The letter cycles every 26 groups and the label grows a character on each
+    wrap, so no two indices share a label. The width has to key off the wrap
+    count -- an earlier version restarted at two letters after ZZ, which handed
+    index 26 the same 'AA' as index 0 and pointed two products at one output
+    folder, where the second silently overwrote the first.
+    """
     idx = max(0, idx)
-    if idx <= 25:
-        c = chr(ord('A') + idx)
-        return c + c
-    n = 2 + (idx - 26) // 26
-    c = chr(ord('A') + ((idx - 26) % 26))
-    return c * n
+    return chr(ord('A') + idx % 26) * (2 + idx // 26)
 
 
 def alpha_pair_to_idx(prefix):
@@ -590,11 +593,13 @@ def freecad_prepare(step_path, freecad_exe, deflection):
         return None
     tmp_dir = tempfile.mkdtemp(prefix="mv_step_")
     obj_out = os.path.join(tmp_dir, os.path.splitext(os.path.basename(step_path))[0] + ".obj")
-    quoted_step = step_path.replace("\\", "\\\\").replace("'", "\'")
-    quoted_obj = obj_out.replace("\\", "\\\\").replace("'", "\'")
+    # Paths go in as %r, not hand-escaped into '...'. A path holding an
+    # apostrophe -- C:\Users\Dan's Files\part.step -- closed the string early
+    # and the script died of a SyntaxError that surfaced as a bare "conversion
+    # failed". repr() handles quotes, backslashes and newlines in one step.
     script = (
         "import Import, MeshPart, FreeCAD\n"
-        "Import.open('%s')\n" % quoted_step +
+        "Import.open(%r)\n" % step_path +
         "doc = FreeCAD.ActiveDocument\n"
         "meshes = []\n"
         "for obj in doc.Objects:\n"
@@ -608,7 +613,7 @@ def freecad_prepare(step_path, freecad_exe, deflection):
         "    m = meshes[0]\n"
         "    for extra in meshes[1:]:\n"
         "        m.addMesh(extra)\n"
-        "    m.write('%s')\n" % quoted_obj
+        "    m.write(%r)\n" % obj_out
     )
     script_path = os.path.join(tmp_dir, "convert.py")
     with open(script_path, "w", encoding="utf-8") as f: