From 503409f189107b305eb291e3281ce00143a125ad Mon Sep 17 00:00:00 2001
From: Moya Chen <72097364+moyapchen@users.noreply.github.com>
Date: Tue, 9 Mar 2021 12:44:05 -0800
Subject: [PATCH] Make CCPE work (#3487)

Cause it's been a while and Google apparently moved the directory and mildly changed the format of the JSON blob.
---
 parlai/tasks/ccpe/agents.py | 6 +-----
 parlai/tasks/ccpe/build.py  | 4 ++--
 2 files changed, 3 insertions(+), 7 deletions(-)

diff --git a/parlai/tasks/ccpe/agents.py b/parlai/tasks/ccpe/agents.py
index 857475c40e6..b7a1b51c68e 100644
--- a/parlai/tasks/ccpe/agents.py
+++ b/parlai/tasks/ccpe/agents.py
@@ -38,15 +38,11 @@ def _setup_data(self):
         fpath = os.path.join(self.opt['datapath'], 'CCPE', 'ccpe.json')
 
         with PathManager.open(fpath, 'r') as infile:
-            data = infile.read()
-            new_data = data.replace('}\n{', '},{')
-            json_data = json.loads(f'[{new_data}]')
+            json_data = json.load(infile)
 
         flattenedData = []
-
         for ep in range(len(json_data)):
             currEp = []
-
             entry = {}
             currSegments = []
             for i, utterance in enumerate(json_data[ep]['utterances']):
diff --git a/parlai/tasks/ccpe/build.py b/parlai/tasks/ccpe/build.py
index dcc0e35c4a1..c743037dbfd 100644
--- a/parlai/tasks/ccpe/build.py
+++ b/parlai/tasks/ccpe/build.py
@@ -10,7 +10,7 @@
 
 RESOURCES = [
     DownloadableFile(
-        'https://storage.googleapis.com/dialog-data-corpus/CCPE-M-2019/data.json',
+        'https://raw.githubusercontent.com/google-research-datasets/ccpe/main/data.json',
         'ccpe.json',
         '4ff051ea7ea60cf0f480c911c7e2cfed56434e2e2c9ea8965ac5e26365773f0a',
         zipped=False,
@@ -20,7 +20,7 @@
 
 def build(opt):
     dpath = os.path.join(opt['datapath'], 'CCPE')
-    version = '1.0'
+    version = '1.1'
 
     if not build_data.built(dpath, version_string=version):
         print('[building data: ' + dpath + ']')