evilsocket / octoghost Goto Github PK

Hi there,

I am just testing out the new 0.5 Ghost and used your script, but I am getting the following error:

$ python octopress2ghost.py octopress_posts ghost_data.json | python -mjson.tool > import-this.json
Traceback (most recent call last):
  File "octopress2ghost.py", line 157, in <module>
    ghost_data['data']['posts'] = posts
KeyError: 'data'
No JSON object could be decoded

The following is Line 154 to Line 159:

ghost_json_file_name = sys.argv[ARG_OUTPUT_JSON]
ghost_data = json.loads(open(ghost_json_file_name).read())

ghost_data['data']['posts'] = posts
ghost_data['data']['tags'] = tags
ghost_data['data']['posts_tags'] = posts_tags

Any ideas?

Hi, sorry to nag you with this but I am unsure why I am getting this error when running the command:

[~/Downloads] python ~/.bin/octopress2ghost.py ~/Projects/Personal/blog/source/_posts GhostData.json | python -mjson.tool > import-this.json
Traceback (most recent call last):
  File "/Users/ericb/.bin/octopress2ghost.py", line 114, in <module>
    value = value.split(':')[0] + ":" + value.split(':')[1]
IndexError: list index out of range
No JSON object could be decoded

Any thoughts? Am I doing it correct?

Almost all my posts were in the .md format instead of .markdown, and that caused problems with the script.

Perhaps could be replaced with simply globbing all files, and leaving it up to the user to ensure only posts are present in the specified folder?

In octropress, categories line can be this:
categories: [Github, git, useful] or sometimes can leave empty categories:.

But script only separate string by blank the_tags = value.split(" "), will cause some error:
Value in [tags.name] cannot be blank. or Duplicate entry found. Multiple values of "xx" found for tags.slug.

I add 3 lines to fix this bug (after line for tag in the_tags:):

tag = tag.strip('[').strip(']').strip(',')
   if tag == "":
       continue

It's not the best solution, but hope help somebody.

I used WP -> Octopress before, and it didn't import all dates. This led to the JSON having no 'created by' which is required by Ghost to import posts.

I wrote a small fix here:

...
                            post_tag_id = post_tag_id + 1
                    else:
                        pass
            else:
                raise Exception('Unexpected exception!')

// Get times if none found
createdAt = post.get('created_at', int(time.mktime(datetime.datetime.strptime(markdown_file[7:17] + ' 00:00', "%Y-%m-%d %H:%M").timetuple()) * 1e3))

    post["created_at"] = createdAt
    post["updated_at"] = createdAt
    post["published_at"] = createdAt

This works due to the fact that filenames themselves have some part of the date - It parses them and assigns the time at 12 o' clock of that day.

It's an ugly solution since I'm pathetic at python, but should work.