Fix url encoding error and update starting instructions

- Some file names have unicode characters. These file names cause problems when they are used to build urls.
- We use ampersand to put the python process in the background, then the process dies when the ssh connection is broken. Therefore, instead, we use setsid so that the process is not tied to the ssh session.

Author: cagdasgerede

Diff for: corrected_change_crawler.py

@@ -43,7 +43,7 @@ def find_interesting_file_paths(self, change_id):
 		paths = []
 		for f in interesting_file_paths:
 			path = self.crawler.build_diff_path(change_id, last_revision_no, f)
-			print path
+			# print path
 			paths.append(path)
 
 		self.logger.info("Successfully done with %s" % change_id)
@@ -68,6 +68,9 @@ def crawl(self, start_change_id, end_change_id):
 			except urllib2.URLError as e:
 				self.logger.error("We failed to reach a server for %s. Reason: %s" % (change_id, e.reason))
 				self.crawler.insert_status(change_id, "ERROR", e.reason)
+			except UnicodeEncodeError as e:
+				self.logger.error("We failed to process a filepath with the unicode format for %s. Reason: %s" % (change_id, e.reason))
+				self.crawler.insert_status(change_id, "ERROR", e.reason)
 			else:
 				self.crawler.insert_status(change_id, "DONE", len(paths))
 				for path in paths:

Diff for: instructions.md

@@ -1,2 +1,3 @@
-python corrected_change_crawler_runner.py --start 300406 --end 300410 --db db/db.sqlite3 --service https://gerrit.wikimedia.org/r
-
+```
+setsid python corrected_change_crawler_runner.py --start 300406 --end 300410 --db db/db.sqlite3 --service https://gerrit.wikimedia.org/r
+```