scripts/convert_sents.py

import srsly
import typer
import warnings
from pathlib import Path
import spacy
from spacy.tokens import DocBin

def convert(lang: str, input_paths: list[Path], output_dir: Path, spans_key: str = "sc"):
    nlp = spacy.blank(lang)
    nlp.add_pipe("sentencizer")
    
    # Ensure output directory exists
    output_dir.mkdir(parents=True, exist_ok=True)

    total_sentences = 0

    # Process each input file
    for input_path in input_paths:
        print(f"Processing file: {input_path}")
        doc_bin = DocBin()
        
        for annotation in srsly.read_jsonl(input_path):
            text = annotation["text"]
            doc = nlp(text)  # Process the document to split into sentences
            
            for sent in doc.sents:
                # Create a new Doc for the sentence
                sent_doc = nlp.make_doc(sent.text)
                spans = []
                for item in annotation["spans"]:
                    # Adjust span start and end for the sentence
                    start = item["start"] - sent.start_char
                    end = item["end"] - sent.start_char
                    label = item["label"]

                    # Only consider spans that are within the sentence
                    if start >= 0 and end <= len(sent.text):
                        span = sent_doc.char_span(start, end, label=label, alignment_mode="contract")
                        if span is None:
                            msg = f"Skipping entity [{start}, {end}, {label}] in the following text because the character span '{sent.text[start:end]}' does not align with token boundaries."
                            warnings.warn(msg)
                        else:
                            spans.append(span)
                
                # Add sentence to DocBin only if it contains spans
                if spans:
                    sent_doc.spans[spans_key] = spans
                    doc_bin.add(sent_doc)
                    total_sentences += 1

        # Write to output file in the specified directory
        output_file = output_dir / f"{input_path.stem}.spacy"
        doc_bin.to_disk(output_file)
    
    print(f"Total sentences with spans: {total_sentences}")

if __name__ == "__main__":
    typer.run(convert)