Fix issue downloading NEET data

Pipfile

@@ -11,6 +11,7 @@ dvc = "*"
 openpyxl = "*"
 papermill = "*"
 ipykernel = "*"
+requests = "*"
 
 [dev-packages]
 

scripts/cpi/dvc.lock

@@ -88,8 +88,8 @@ stages:
       size: 8893
     - path: working/metadata.csv
       hash: md5
-      md5: 4ec0da4994782d25aca721561815e260
-      size: 7650
+      md5: 0da329f21f3d8a08af2216103729f224
+      size: 7642
     outs:
     - path: src/_data/sources/cpi/cpi_barchart.csv
       hash: md5

scripts/labour-market/dvc.lock

@@ -13,8 +13,8 @@ stages:
       size: 608
     - path: working/LMS_data.csv
       hash: md5
-      md5: 87e8d61a9a784f9db2a56b1a67bccb8f
-      size: 16991591
+      md5: 0fa1500fa1ad5467cf0710943f012825
+      size: 16991592
     outs:
     - path: data/labour-market/monthly-rolling.csv
       hash: md5
@@ -29,13 +29,13 @@ stages:
       size: 359985
     - path: ../../scripts/util/
       hash: md5
-      md5: 13221061b7639bdb76f26c9ffb1fc256.dir
-      size: 13180
+      md5: bbd9c17bdb21207e9126ae17bc04e2fc.dir
+      size: 13118
       nfiles: 9
     - path: ../../working/metadata.csv
       hash: md5
-      md5: 4ec0da4994782d25aca721561815e260
-      size: 7650
+      md5: 0da329f21f3d8a08af2216103729f224
+      size: 7642
     - path: config.py
       hash: md5
       md5: c9187b99fb9c48a24be707561d15eaa2

scripts/util/downloader.py

@@ -1,8 +1,7 @@
-from urllib.request import build_opener, install_opener, urlretrieve
+import requests
 
 
 def download_file(url, filename, headers=[]):
-    opener = build_opener()
-    opener.addheaders = headers
-    install_opener(opener)
-    urlretrieve(url, filename)
+    response = requests.get(url)
+    with open(filename, 'wb') as file:
+        file.write(response.content)

scripts/vacancies/dvc.lock

@@ -13,8 +13,8 @@ stages:
       size: 5733
     - path: working/metadata.csv
       hash: md5
-      md5: 4ec0da4994782d25aca721561815e260
-      size: 7650
+      md5: 0da329f21f3d8a08af2216103729f224
+      size: 7642
     - path: working/vacancies/vacancies_by_sector.csv
       hash: md5
       md5: d87aa1a94f87db2274f2b87c9ee0d9ab
@@ -44,8 +44,8 @@ stages:
       size: 822
     - path: working/LMS_data.csv
       hash: md5
-      md5: 87e8d61a9a784f9db2a56b1a67bccb8f
-      size: 16991591
+      md5: 0fa1500fa1ad5467cf0710943f012825
+      size: 16991592
     - path: working/lookups/LMS_variable_lookup.csv
       md5: faf1d16226f95f8d2448b90d1d5868a0
       size: 137661

working/LMS_data.csv.dvc

@@ -1,12 +1,12 @@
-md5: ff0510b134124df1867a91a9882bc780
+md5: d7f482f9d61f988b2c308c01a34be96a
 deps:
-- checksum: '"fa83c30aceea0bbe5a6d6427acef0b8d3bc53c483821062d7cf889dadf00517b"'
-  size: 16991591
+- checksum: '"c699a3037cbb8a20a37a025397a3d530eaa8ff7be0578333bcd51f2d3355404d"'
+  size: 16991592
   path: 
     https://github.com/economic-analytics/edd/blob/main/data/csv/LMS_data.csv?raw=true
   hash: md5
 outs:
-- md5: 87e8d61a9a784f9db2a56b1a67bccb8f
-  size: 16991591
+- md5: 0fa1500fa1ad5467cf0710943f012825
+  size: 16991592
   path: LMS_data.csv
   hash: md5

working/metadata.csv.dvc

@@ -1,13 +1,13 @@
-md5: f92e654b731dfc09c164adf560421856
+md5: 6e55818c9d84fc8f1da2c5ad320405d8
 # frozen: true
 deps:
-- checksum: '"3e601232be60001167fd43759b008e7b6bc434ae603a7ae628fee84867623ac4"'
-  size: 7650
+- checksum: '"b3fbe22876b819ad38760dc198b67328657e026030573b7aab6b5ece9441900f"'
+  size: 7642
   hash: md5
   path: 
     https://raw.githubusercontent.com/economic-analytics/edd/main/data-raw/edd_dict.csv
 outs:
-- md5: 4ec0da4994782d25aca721561815e260
-  size: 7650
+- md5: 0da329f21f3d8a08af2216103729f224
+  size: 7642
   hash: md5
   path: metadata.csv