Fix off-by-one in CodeCopyrightTagger end position (missing newline)

[Chessing234]

Bug

CodeCopyrightTagger._extract_copyright_spans in python/dolma/taggers/code/code_taggers.py produces comment_block spans whose end is short by one per non-empty comment line. The last comment line of a leading copyright block ends up truncated mid-line.

Root cause

```python
lines = text.split("\n")
...
end = 0
for line in lines:
if line.startswith("//") or line.startswith("#") or line.startswith("--") or not line:
skip = skip + 1
if not line:
end += 1
else:
end += len(line)
else:
break
```

str.split(\"\\n\") drops the newline separators, so each matched line occupied len(line) + 1 bytes in the original text (its content plus the trailing \"\\n\"). The empty-line branch correctly adds 1 (empty content + the newline), but the non-empty branch adds only len(line), forgetting the newline.

Worked example

`text = "// a\n// b\nother"`:

• lines = [\"// a\", \"// b\", \"other\"]
• Current: end = 4 + 4 = 8. The emitted span is [0, 8) = \"// a\\n// \", cutting off \"b\\n\".
• Expected: end = 5 + 5 = 10, which is exactly where \"other\" begins.

This downstream feeds Span(start=0, end=end, type=\"comment_block\", ...) and _score, so filters and downstream taggers see a slightly-too-small copyright region.

Fix

Collapse both branches to end += len(line) + 1. For empty lines this is 0 + 1 = 1, matching the prior behavior. For non-empty comment lines it now correctly includes the dropped newline. Two lines added, four removed; the if not line / else split is no longer needed.

```diff

•            if not line:
  

•                end += 1
  

•            else:
  

•                end += len(line)
  

•            # +1 accounts for the \"\\n\" separator that str.split dropped
  

•            end += len(line) + 1
  

```