Changelog conversion script improvements and bugfix (#102)

Author: endorama

tools/asciidoc_to_fragments.py

@@ -8,7 +8,24 @@
 from os import makedirs
 from os.path import expanduser
 from datetime import datetime
-
+from hashlib import sha1
+
+# Using this script
+# Run it from destination repository root with:
+# python /path/to/elastic-agent-changelog-tool/tools/asciidoc_to_fragments.py --path CHANGELOG.next.asciidoc --workdir $PWD
+#
+# If errors arise you should at first try to solve them in the source changelog, 
+# so that if you re-run the script you are not required to apply the same fixes
+# again.
+# Fixable errors:
+# - look for duplicated entries
+# - no response from Github API: look for missing or wrong data (es issue number instead of PR number) in {pull}
+# - no PR/issue fields: no {pull} or {issue field present}
+# - multiple PRs/issues found: the tool does not support multiple {pull} or {issue} on the same line; split them or remove all but one {issue} and one {pull}
+# - issue info lost due to multiple repositories: remove the one referring to an external repository
+# - look for files starting with "1000000*", as this timestamp means something is wrong (missing {pull} maybe?)
+#
+# For the remaining errors, fix them in the created fragments.
 
 api_url = "https://api.github.com/repos/"
 github_token_location = "/.elastic/github.token"
@@ -30,20 +47,26 @@
 kind_token = "===="
 field_token = "-"
 
-def write_fragment(title, fragment_timestamp, fragment_dict):
+def write_fragment(filename, fragment_timestamp, fragment_dict):
     if not fragment_timestamp:
         fragment_timestamp = str(1000000000 + fragments_counter)
 
     path = "".join([fragments_path,
                     fragment_timestamp,
                     "-",
-                    title,
+                    filename,
                     ".yaml"])
 
     with open(path, 'w+') as f:
         for k, v in fragment_dict.items():
             f.write(f"{k}: {v}\n")
 
+    # print path and SHA1 of it's content, for verification purposes
+    with open(path, 'r') as f:
+        content = f.read()
+        hash_object = sha1(content.encode('utf-8'))
+        print(path, hash_object.hexdigest())
+
 def get_event_timestamp(repository, event, number):
     token_path = ''.join([expanduser("~"), github_token_location])
     with open(token_path, 'r') as f:
@@ -106,10 +129,8 @@ def parse_line(line, kind):
             issue_number, issue_repo = number, repo_link
 
     if pr_repo:
-        fragment_dict["repository"] = pr_repo
         fragment_timestamp = get_event_timestamp(pr_repo, "pulls", pr_number)
     elif issue_repo:
-        fragment_dict["repository"] = issue_repo
         fragment_timestamp = get_event_timestamp(issue_repo, "issues", issue_number)
 
     if fragment_timestamp == "not_found":