JsonCorpusToTSV

Conversion/ToTSV/JsonCorpusToTSV/JsonCorpusToCsv.py

+#######
+# jsonCorpusToCsv.py
+# description : change a json GarganText corpus into a csv legacy corpus
+# licence : AGPL + CECILL v3
+# author : quentin lobbé - qlobbe@iscpif.fr
+#######
+
+# python3 jsonCorpusToCsv.py corpus.json
+
+import sys
+import csv
+import json
+from zipfile import ZipFile
+
+try :
+    pathCorpus = sys.argv[1]
+except :
+    print ("! args error\n")
+    sys.exit(0)
+	
+def readZipFile(path):
+	with ZipFile(path, 'r') as f:
+		file = f.open(f.namelist()[0])
+		return json.load(file)
+
+def readJson(path) :
+	file = open(path)
+	return json.load(file)
+
+if pathCorpus.split('.')[1] == 'zip':
+	corpusJson = readZipFile(pathCorpus)
+else:
+	corpusJson = readJson(pathCorpus)
+
+output = open(str(pathCorpus.split('.')[0]) + ".csv", "w")
+header = "title\tsource\tpublication_year\tpublication_month\tpublication_day\tabstract\tauthors\tweight\n"
+output.write(header)
+
+for row in corpusJson['corpus'] :
+	doc = row['document']['hyperdata']
+	abstract = "empty"
+	authors  = "empty"
+	title  = "empty"
+	source = "empty"
+
+	if 'title' in doc.keys() :
+		title = doc['title'].replace('"','').replace('\t','')
+
+	if 'source' in doc.keys() :
+		source = doc['source'].replace('"','').replace('\t','')
+
+	if 'abstract' in doc.keys() :
+		abstract = doc['abstract'].replace('"','').replace('\t','')
+
+
+	if 'authors' in doc.keys() :
+		authors = doc['authors']
+
+
+	output_row = title + "\t" + source + "\t" + str(doc['publication_year']) + "\t" + str(doc['publication_month']) + "\t" + str(doc['publication_day']) + "\t" + abstract + "\t" + authors + "\t" + str(1) + "\n"
+
+	output.write(output_row)
\ No newline at end of file

Conversion/ToTSV/JsonCorpusToTSV/README.md

+# JsonCorpusToTSV
+
+## About The project
+
+JsonCorpusToTSv transform a JsonCorpus from Gargantext into a TSV corpus.
+
+## Usage
+```shell
+python3 JsonCorpusToTSV.py corpus.json
+```
+
+corpus.json -> GarganText corpus in json format
+
+Output a TSV legacy corpus : corpus.csv
+
+You can also use a zip file with a json corpus in it
+```shell
+python3 JsonCorpusToTSV.py corpus.zip
+```
\ No newline at end of file