chore: bump version to 0.9.2

[percevalw]

Changed

• Default to fp16 when inferring with gpu
• Support inputs parameter in TrainablePipe.postprocess(...) method (as in edsnlp)
• We now check that the user isn't trying to write a single file in a split fashion (when write_in_worker is True or num_rows_per_file is not None) and raise an error if they do

Fixed

• Batches full of empty content boxes no longer crash the huggingface-embedding component
• Ensure models are always loaded in non training mode
• Improved performance of edspdf.data methods over a filesystem (fs parameter)

[sonarcloud[bot]]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[github-actions[bot]]

Coverage Report

Name	Stmts	Miss	∆ Miss	Cover
TOTAL	3134	76	0	97.57%

Files without new missing coverage

Name	Stmts	Miss	∆ Miss	Cover
edspdf/visualization/annotations.py Was already missing at line 67 elif isinstance(colors, list): - colors = {label: color for label, color in zip(unique_labels, colors)}	32	1	0	96.88%
edspdf/utils/optimization.py Was already missing at line 29 def param_groups(self, value): - self.optim.param_groups = value Was already missing at line 33 def state(self): - return self.optim.state Was already missing at line 37 def state(self, value): - self.optim.state = value	68	3	0	95.59%
edspdf/utils/lazy_module.py Was already missing at line 92 """ - return __all__	31	1	0	96.77%
edspdf/utils/file_system.py Was already missing at line 26 ) -> list: - return [ os.path.join(dirpath, f)	24	1	0	95.83%
edspdf/utils/collections.py Was already missing at line 149 def __getstate__(self): - return {"seq": self.seq} Was already missing at lines 152-154 def __setstate__(self, state): - self.seq = state["seq"] - self.flatten = None Was already missing at lines 252-254 base[attr] = val - except (KeyError, TypeError): - setattr(base, attr, val) return base	166	5	0	96.99%
edspdf/utils/alignment.py Was already missing at line 19 if len(src_boxes) == 0 or len(dst_boxes) == 0: - return []	30	1	0	96.67%
edspdf/trainable_pipe.py Was already missing at line 76 if cache_key in cache: - return cache[cache_key] res = fn(self, doc) Was already missing at lines 349-357 """ - batch = [ ... - return batch Was already missing at line 380 if hasattr(self, "compiled"): - res = self.compiled(batch) else: Was already missing at line 459 if pipe_overrides: - overrides[name] = pipe_overrides tensor_dict = {	202	8	0	96.04%
edspdf/structures.py Was already missing at line 186 def page(self): - return next(p for p in self.doc.pages if p.page_num == self.page_num) Was already missing at line 192 if self_page_num < other_page_num: - return True if self_page_num > other_page_num: Was already missing at line 194 if self_page_num > other_page_num: - return False Was already missing at line 222 - return ((self.y0 + self.y1) / 2, (self.x0 + self.x1) / 2) < ( (other.y0 + other.y1) / 2, Was already missing at line 250 def __str__(self): - return self.text	97	5	0	94.85%
edspdf/registry.py Was already missing at lines 112-114 raise - except ConfitValidationError as e: - errors.append(e.raw_errors) if not errors:	90	2	0	97.78%
edspdf/processing/utils.py Was already missing at line 15 if isinstance(res, types.GeneratorType): - results.extend(res) else: Was already missing at lines 66-78 ) -> Iterable[List[T]]: - batch = [] ... - yield batch	55	13	0	76.36%
edspdf/processing/simple.py Was already missing at lines 27-29 no_grad = sys.modules["torch"].no_grad - except (KeyError, AttributeError): - no_grad = nullcontext reader = lc.reader	51	2	0	96.08%
edspdf/processing/multiprocessing.py Was already missing at lines 226-230 if os.environ.get("TORCH_SHARING_STRATEGY"): - try: - torch.multiprocessing.set_sharing_strategy(os.environ["TORCH_SHARING_STRATEGY"]) - except NameError: - pass Was already missing at line 248 def save_align_devices_hook(pickler: Any, obj: Any): - pickler.save_reduce(load_align_devices_hook, (obj.__dict__,), obj=obj) Was already missing at lines 251-258 def load_align_devices_hook(state): - state["execution_device"] = MAP_LOCATION ... - AlignDevicesHook = None Was already missing at line 416 if lc.sort_chunks: - docs.sort( key=doc_size_fns.get( Was already missing at line 456 - new_batch_iterator = None Was already missing at line 495 if task is None and stage == self.exchanger.num_stages + 1: - return # Non prioritized STOP signal: there are no more tasks to process Was already missing at lines 577-579 else: - batch = gpu_pipe.prepare_batch(docs, device=device) - inputs = None active_batches[batch_id] = (docs, task_id, inputs) Was already missing at line 1016 if v is not None: - os.environ[k] = v	406	17	0	95.81%
edspdf/pipes/extractors/pdfminer.py Was already missing at line 161 if len(text) == 0: - continue content_boxes.append( Was already missing at line 222 else: - fontname, italic, bold = (None, None, None) else:	88	2	0	97.73%
edspdf/pipeline.py Was already missing at line 992 if overrides: - config = config.merge(overrides) pwd = os.getcwd()	335	1	0	99.70%
edspdf/lazy_collection.py Was already missing at lines 323-326 """Moves the pipeline to a given device""" - for name, pipe, *_ in self.torch_components(): - pipe.to(device) - return self	120	3	0	97.50%
edspdf/layers/relative_attention.py Was already missing at lines 157-159 if head_size is None and key_size is not None: - assert key_size % n_heads == 0 - head_size = key_size // n_heads value_head_size = None Was already missing at line 172 ): - self.register_buffer("position_embedding", position_embedding) else: Was already missing at line 177 if same_key_query_proj: - self.content_query_proj = self.content_key_proj else: Was already missing at line 196 if same_key_query_proj or same_positional_key_query_proj: - self.position_query_proj = self.position_key_proj else: Was already missing at line 355 if mask.ndim == 3: - mask = mask[:, :, :, None] Was already missing at line 366 - return attn	117	7	0	94.02%
edspdf/data/parquet.py Was already missing at line 49 # read in worker -> each task is a non yet parsed line - return ( (line, 1)	107	1	0	99.07%
edspdf/data/pandas.py Was already missing at line 101 if isinstance(rec, dict): - rec.pop(FILENAME, None) return records, len(records)	44	1	0	97.73%
edspdf/data/files.py Was already missing at lines 100-102 if self.load_annotations and self.filesystem.exists(json_path): - with self.filesystem.open(json_path) as f: - record["annotations"] = json.load(f)	102	2	0	98.04%

35 files skipped due to complete coverage.

Coverage success: total of 97.57% is above 97.57% 🎉