From 99a80215c183f8c8467403f0e6ceffa628b2a756 Mon Sep 17 00:00:00 2001
From: Shashank Mittal <shashank.mittal.mec22@itbhu.ac.in>
Date: Wed, 28 Feb 2024 00:39:15 +0530
Subject: [PATCH] added common funcs to utils.py

Signed-off-by: Shashank Mittal <shashank.mittal.mec22@itbhu.ac.in>
---
 .../languages/English/nouns/format_nouns.py   | 40 ++-----------------
 src/scribe_data/utils.py                      | 34 ++++++++++++++++
 2 files changed, 38 insertions(+), 36 deletions(-)

diff --git a/src/scribe_data/extract_transform/languages/English/nouns/format_nouns.py b/src/scribe_data/extract_transform/languages/English/nouns/format_nouns.py
index 49b8ac07f..0e31f4970 100644
--- a/src/scribe_data/extract_transform/languages/English/nouns/format_nouns.py
+++ b/src/scribe_data/extract_transform/languages/English/nouns/format_nouns.py
@@ -6,31 +6,16 @@
 """
 
 import collections
-import json
-import os
 import sys
 
+from scribe_data.utils import export_formatted_data, load_queried_data
+
 LANGUAGE = "English"
 QUERIED_DATA_TYPE = "nouns"
-QUERIED_DATA_FILE = f"{QUERIED_DATA_TYPE}_queried.json"
-PATH_TO_SCRIBE_ORG = os.path.dirname(sys.path[0]).split("Scribe-Data")[0]
-LANGUAGES_DIR_PATH = (
-    f"{PATH_TO_SCRIBE_ORG}/Scribe-Data/src/scribe_data/extract_transform/languages"
-)
 
 file_path = sys.argv[0]
 
-update_data_in_use = False  # check if update_data.py is being used
-if f"languages/{LANGUAGE}/{QUERIED_DATA_TYPE}/" not in file_path:
-    data_path = QUERIED_DATA_FILE
-else:
-    update_data_in_use = True
-    data_path = (
-        f"{LANGUAGES_DIR_PATH}/{LANGUAGE}/{QUERIED_DATA_TYPE}/{QUERIED_DATA_FILE}"
-    )
-
-with open(data_path, encoding="utf-8") as f:
-    nouns_list = json.load(f)
+nouns_list, update_data_in_use = load_queried_data(LANGUAGE, QUERIED_DATA_TYPE, file_path)
 
 nouns_formatted = {}
 
@@ -94,21 +79,4 @@
 
 nouns_formatted = collections.OrderedDict(sorted(nouns_formatted.items()))
 
-export_path = f"../formatted_data/{QUERIED_DATA_TYPE}.json"
-if update_data_in_use:
-    export_path = (
-        f"{LANGUAGES_DIR_PATH}/{LANGUAGE}/formatted_data/{QUERIED_DATA_TYPE}.json"
-    )
-
-with open(
-    export_path,
-    "w",
-    encoding="utf-8",
-) as file:
-    json.dump(nouns_formatted, file, ensure_ascii=False, indent=0)
-
-print(
-    f"Wrote file {QUERIED_DATA_TYPE}.json with {len(nouns_formatted):,} {QUERIED_DATA_TYPE}."
-)
-
-os.remove(data_path)
+export_formatted_data(LANGUAGE, QUERIED_DATA_TYPE, nouns_formatted, update_data_in_use)
diff --git a/src/scribe_data/utils.py b/src/scribe_data/utils.py
index c6b894303..fbaa98e4c 100644
--- a/src/scribe_data/utils.py
+++ b/src/scribe_data/utils.py
@@ -11,6 +11,9 @@
     get_language_words_to_remove,
     get_language_words_to_ignore,
     get_path_from_format_file,
+    get_language_dir_path,
+    load_queried_data,
+    export_formatted_data,
     get_path_from_load_dir,
     get_path_from_et_dir,
     get_ios_data_path,
@@ -22,6 +25,7 @@
 
 import ast
 import json
+import os
 import sys
 from importlib import resources
 from pathlib import Path
@@ -240,6 +244,36 @@ def get_language_words_to_ignore(language: str) -> list[str]:
     )
 
 
+def get_language_dir_path(language):
+    PATH_TO_SCRIBE_ORG = os.path.dirname(sys.path[0]).split("Scribe-Data")[0]
+    return f"{PATH_TO_SCRIBE_ORG}/Scribe-Data/src/scribe_data/extract_transform/languages/{language}"
+
+
+def load_queried_data(language, data_type, file_path):
+    queried_data_file = f"{data_type}_queried.json"
+    update_data_in_use = False
+
+    if f"languages/{language}/{data_type}/" not in file_path:
+        data_path = queried_data_file
+    else:
+        update_data_in_use = True
+        data_path = f"{get_language_dir_path(language)}/{data_type}/{queried_data_file}"
+
+    with open(data_path, encoding="utf-8") as f:
+        return json.load(f), update_data_in_use
+
+
+def export_formatted_data(language, data_type, formatted_data, update_data_in_use):
+    if update_data_in_use:
+        export_path = f"{get_language_dir_path(language)}/formatted_data/{data_type}.json"
+    else:
+        export_path = f"{data_type}.json"
+
+    with open(export_path, "w", encoding="utf-8") as file:
+        json.dump(formatted_data, file, ensure_ascii=False, indent=0)
+    print(f"Wrote file {data_type}.json with {len(formatted_data):,} {data_type}.")
+
+
 def get_path_from_format_file() -> str:
     """
     Returns the directory path from a data formatting file to scribe-org.