64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178 | def apply_fixes(
bib_path: str | Path,
results: list[dict[str, Any]],
output_path: str | Path | None = None,
fields: list[str] | None = None,
dry_run: bool = False,
) -> tuple[list[dict[str, Any]], set[str]]:
"""Rewrite mismatch entries in *bib_path* with InspireHEP canonical values.
Only entries with ``status == "mismatch"`` that have an ``inspire_record``
are modified. Entries with ``status == "missing"`` are left untouched
(we don't know which record to fix them to).
The output file is written with clean entries first and still-flagged
entries (missing + mismatches with no fixable InspireHEP record) moved
to the end, separated by a comment block.
Parameters
----------
bib_path : str or Path
Original ``.bib`` file to read.
results : list[dict[str, Any]]
Parsed contents of ``results.json`` produced by Step 1.
output_path : str or Path, optional
Where to write the fixed bib. Defaults to a new file named
``<stem>_fixed.bib`` next to *bib_path*.
When *dry_run* is ``True`` nothing is written regardless.
fields : list[str], optional
Restrict which fields are updated. Defaults to
:data:`DEFAULT_FIX_FIELDS`.
dry_run : bool, optional
When ``True``, compute changes but do not write anything.
Returns
-------
applied : list[dict[str, Any]]
One entry per modified field::
{"key": str, "field": str, "old": str, "new": str}
still_flagged : set[str]
Citation keys that still need manual attention after the fix
(missing entries + mismatches with no fixable InspireHEP record).
Raises
------
FileNotFoundError
Raised if *bib_path* does not exist.
"""
bib_path = Path(bib_path)
if not bib_path.exists():
raise FileNotFoundError(f"Bib file not found: {bib_path}")
if fields is None:
fields = DEFAULT_FIX_FIELDS
if output_path is None:
output_path = bib_path.with_name(bib_path.stem + "_fixed" + bib_path.suffix)
output_path = Path(output_path)
# Build a map of key → canonical updates for mismatch entries only.
fix_map: dict[str, dict[str, str]] = {}
for r in results:
if r.get("status") != "mismatch":
continue
updates = _canonical_values(r, fields)
if updates:
fix_map[r["key"]] = updates
# Keys that remain problematic after the fix pass:
# - all missing entries (no record to fix to)
# - mismatch entries where we found no fixable canonical values
still_flagged: set[str] = set()
for r in results:
key = r["key"]
if r.get("status") == "missing" or key not in fix_map:
still_flagged.add(key)
library = bibtexparser.parse_file(str(bib_path))
applied: list[dict[str, Any]] = []
for entry in library.entries:
updates = fix_map.get(entry.key)
if not updates:
continue
current = {f.key: f.value for f in entry.fields}
for field_name, new_value in updates.items():
old_value = current.get(field_name, "")
if old_value == new_value:
continue
applied.append(
{"key": entry.key, "field": field_name, "old": old_value, "new": new_value}
)
if not dry_run:
entry.set_field(Field(field_name, new_value))
if not dry_run:
# Write a patched intermediate file, then reformat to move
# still-flagged entries to the end with a separator comment.
import tempfile
with tempfile.NamedTemporaryFile(
mode="w", suffix=".bib", delete=False, encoding="utf-8"
) as tmp:
tmp.write(bibtexparser.write_string(library))
tmp_path = Path(tmp.name)
try:
write_reformatted_bib(tmp_path, still_flagged, output_path)
finally:
tmp_path.unlink(missing_ok=True)
return applied, still_flagged
|