json_gz2nt_gz.py

# json_gz2nt_gz.py

import json
import gzip
import sys

from datetime import datetime

# Define the filename of the gzipped file (replace with your actual path)

def read_gz_by_row(FILENAME):
    try:
    # Open the file in 'rb' (read binary) mode with gzip decompression
        with gzip.open(FILENAME, 'rb') as f:
            # Iterate through each line in the file
            for line in f:
            # Decode the line from bytes to string (adjust encoding if needed)
                decoded_line = line.decode('utf-8')
                yield decoded_line.strip()
    
    # Process the decoded line as needed
        print(decoded_line.rstrip())  # Print the line with trailing whitespace removed
    except FileNotFoundError:
        print(f"Error: File '{FILENAME}' not found.")

def limit_string_size(text, max_size=1024, suffix="..."):
  """Limits the size of a string to a specified maximum, adding an optional suffix.

  Args:
      text: The string to limit.
      max_size: The maximum allowed size (default: 1024 characters).
      suffix: The suffix to add if the string is truncated (default: "...").

  Returns:
      The truncated string if it exceeds the limit, otherwise the original string.
  """

  if len(text) > max_size:
    # Truncate at the character boundary closest to max_size
    truncated_text = text[:max_size - len(suffix)]  # Exclude suffix length
    return truncated_text + suffix  # Add the suffix after truncation

  else:
    return text  # Return original string if it's within the limit



def fix_string(string):

    string = string.replace('\\','')
    string = string.replace("'","")
    string = string.replace('"','')
    string = string.replace("\n",'')
    string = limit_string_size(string)

    return string


def object2nt(json_doc):

    ntriples_str = ''
    
    for key, value in json_doc.items():

        #print(key, type(value), value)

        if not value is None:
            if isinstance(value, bool):
                if value:
                    ntriple = f"<http://bio2rdf.org/{NAMESPACE}:{doc_id}> <http://bio2rdf.org/intermine_voc:{key}> true .\n"
                else:
                    ntriple = f"<http://bio2rdf.org/{NAMESPACE}:{doc_id}> <http://bio2rdf.org/intermine_voc:{key}> false .\n"
            elif isinstance(value, str):
                value = fix_string(value)
                ntriple = f"<http://bio2rdf.org/{NAMESPACE}:{doc_id}> <http://bio2rdf.org/intermine_voc:{key}> '{value}' .\n"
                if key == 'class_name':
                    ntriple1 = f"<http://bio2rdf.org/{NAMESPACE}:{doc_id}> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://bio2rdf.org/intermine_voc:{value}> .\n"
                    ntriple = ntriple1 + ntriple
            else:
                ntriple = f"<http://bio2rdf.org/{NAMESPACE}:{doc_id}> <http://bio2rdf.org/intermine_voc:{key}> {value} .\n"
        
            #print(ntriple)
            ntriples_str = ntriples_str + ntriple
    return ntriples_str


def relation2nt(json_doc):

    #print(json_doc, type(json_doc))

    ntriple1 = f"<http://bio2rdf.org/{json_doc['namespace']}:{json_doc['class_name_from']}:{str(json_doc['id_from'])}> "
    ntriple2 = f"<http://bio2rdf.org/intermine_voc:{json_doc['predicate'].replace('.id','_id')}> "
    ntriple3 = f"<http://bio2rdf.org/{json_doc['namespace']}:{json_doc['class_name_to']}:{str(json_doc['id_to'])}> .\n"

    ntriple = ntriple1 + ntriple2 + ntriple3
    
    return ntriple



def ndjson2nt(TYPE, NAMESPACE, FOLDER):
    file_nt_gz = gzip.open(f"{FOLDER}/{NAMESPACE}-{TYPE}.nt.gz2", "wb")

    ctr = 0
    str_len = 0
    
    
    for json_str in read_gz_by_row(f"{FOLDER}/{NAMESPACE}-{TYPE}.ndjson.gz"):
        ctr += 1
        if ctr % 100000 == 0:
            print(NAMESPACE, TYPE, datetime.utcnow().isoformat(), ctr, str_len)

        json_dict = json.loads(json_str)
        json_doc = json_dict['_source']
        doc_id = json_doc['doc_id'].replace(NAMESPACE+'-','').replace('-',':')
        #print(ctr, json_doc['class_name'], doc_id)
         
        if TYPE == 'object':
            ntriple_str = object2nt(json_doc)
        else:
            ntriple_str = relation2nt(json_doc)

        str_len = str_len + len(ntriple_str)
        
        file_nt_gz.write(ntriple_str.encode("utf-8"))

    file_nt_gz.close()
    print(NAMESPACE, TYPE, datetime.utcnow().isoformat(), ctr, str_len)

# python3 json_gz2nt_gz.py relation wormmine ../data
# python3 json_gz2nt_gz.py relation yeastmine ../data

NAMESPACE = 'flymine'
FOLDER = './'
# relation or object
TYPE = 'relation'
TYPE = 'object'
TYPE = sys.argv[1]
NAMESPACE = sys.argv[2]
FOLDER = sys.argv[3]

try:
    ndjson2nt(TYPE, NAMESPACE, FOLDER)
except KeyboardInterrupt:
    print("Stopping execution due to Ctrl+C...")