pelican
pelican copied to clipboard
getpelican
Pelican should print a warning when encountering a .md file while markdown is not installed
Hello,
Currently, pelican will skip .md files if the "markdown" python packge is not installed. This result in a WARNING: No valid files found in content." which is extremely confusing and has made me lost 30min of debugging.
A better behavior would be to print a warning because there isn't a situation where a user would add a .md file to his content/ while not wanting it to be parsed.
Kind regards,
Well, the python adage as always been "explicit is better than implicit", in this case you have no indication that there might be a problem, it just doesn't work, it's a bit like saying "oh you need to set up this environment variable for this to be working", it wipes out the possibility of self discovery which is, for me, a bad practice, documented or not.
@Psycojoker if you need 30min to look up the documentation and find a solution then that is a documentation problem. However I think it is reasonably good at explaining the issue.
As for more warnings regarding the content when no Markdown Parser is available: Currently pelican is only looking for files of certain extensions when their reader is activated. If no markdown is available, .md files are not even looked at and are 'invisible' to pelican. I think that is great, it is one of the few occasions where pelican is actually abstracted well and can easily be extended or modified to only use a subset of readers that is needed.
Your suggestion would have implications for everyone who does not want to use markdown. Actually scanning the content directory for .md files when just using .rst could be a huge time waste for large sites with lots of content files.
A better alternative could be to print the active Readers and their supported file endings when starting up pelican, but again, this is probably not information you want to see on every run and would rather hide it to the debug menu in case 'no valid content' warning is not triggered.
Your usage of fat formatting and the way you phrase your comments makes you sound like a prig. When I first read it I wanted to react similarly to @avaris.
A better alternative could be to print the active Readers and their supported file endings when starting up pelican, but again, this is probably not information you want to see on every run and would rather hide it to the debug menu in case 'no valid content' warning is not triggered.
If find the "only active when 'no valid content' case is encounter" to be a good middle ground. In my head you were using something like os.walk which would have allowed this operation without to much overhead since you would have been scanning all file names.
Your usage of fat formatting and the way you phrase your comments makes you sound like a prig. When I first read it I wanted to react similarly to @avaris.
Sorry, the opposite was actually intent, as in "this is only my opinion which my not be the best opinion", I should have made it clearer to avoid ambiguity, my apologies :/
We use os.walk to scan files, but we filter it to the ones pelican can process. That is obtained from the active readers. And missing dependencies isn't the only way to disable a reader (see READERS config).
Also, people might put .md files in their content simply as a static file. Just the existence of an .md file isn't as clear an indication as it might seem.
Having said that, listing active readers in debug log is a reasonable idea.
What about replacing "WARNING: No valid files found in content" by "WARNING: No valid files found in content (did you forget to install a language support plugin?)"
There appears to be consensus on @cpitclaudel's suggestion and it looks like a pretty straightforward change.
What about replacing "WARNING: No valid files found in content" by "WARNING: No valid files found in content (did you forget to install a language support plugin?)"
Plus maybe a link or a hint to the part in the documentation? I have to google the error everytime, come to such a thread, and figure the problem. Specially since many people including me, think of pelican as something which handles .md files by default.
@adrinjalali: This is mentioned over and over again in the docs. I'm fine with adding more descriptive error-handling, but IMHO this is not a documentation problem.
@justinmayer: oh yeah absolutely. I agree that the documentation couldn't be more explicit. I'm also talking about the warning message.
Ah, re-reading your comment, I now understand what you meant. Apologies for the misunderstanding. 😅
The problem with the docs is that I if I see "no content found", my assumption is not that I should be looking at the docs about which directories are searched, or that I've put things in the wrong directories.
How about "no restructedText content found", where we make a list of the language parsers we ran. Then I would see that it isn't even looking for markdown.
OK, we can elaborate the warning. How about:
WARNING: No valid files found in content for the active readers:
| RstReader (.rst)
| HTMLReader (.html, .htm)
...
Looks perfect to me (thumbsup)
On Sun, May 7, 2017, 17:52 Deniz Turgut, [email protected] wrote:
OK, we can elaborate the warning. How about:
WARNING: No valid files found in content for the active readers: | RstReader (.rst) | HTMLReader (.html, .htm) ...
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/getpelican/pelican/issues/1868#issuecomment-299715188, or mute the thread https://github.com/notifications/unsubscribe-auth/ABljeGgwY8LL2w29lpQbUKjkKlKV_Bqjks5r3ejSgaJpZM4GlCZR .
Would anyone in this thread like to help out by implementing the suggestion @avaris proposed?
You mean something like this commit @justinmayer ? https://github.com/adrinjalali/pelican/commit/be4e04f774f779965e3fefea21995a4c9337af4e
example output:
WARNING: No valid files found in content for the active readers:
WARNING: | BaseReader (static)
WARNING: | RstReader (rst)
WARNING: | HTMLReader (htm, html)
@adrinjalali: Perhaps a single WARNING in the output, as @avaris suggested above?
Changed the warning, now it shows:
WARNING: No valid files found in content for the active readers:
| HTMLReader (htm, html)
| BaseReader (static)
| RstReader (rst)
Although, I'd prefer this implementation, up to you.
https://github.com/adrinjalali/pelican/compare/master...adrinjalali:issue89
I realize this issue is closed & the above has been merged. I just ran into this situation and ended up solving it by coming here and reading the thread - despite the error message and the documentation, etc. So I could have done more, for sure, but I still think the error processing / message could be clearer. I'd suggest looking at the elements of the MARKUP list and trying an import based on what's in that list. That way you could issue an error like "You have listed 'md' in your MARKUP variable but you don't have the markdown module installed." and maybe even say something like "Did you forget to run pip install markdown?"
There's a bit of a "blame-the-user" feel to this thread...
Ah... sorry! I guess that was something used in older versions and that has been removed. In my pelicanconf.py I have
MARKUP = ('md', 'ipynb')
If it was still a thing (or maybe it has been renamed?) I was thinking it could provide a way to avoid walking the filesystem looking for (in this case) .md files, etc. Thanks @avaris & sorry for the noise!
Yeah, there is no setting like that. Extensions to process are obtained from active readers.
BTW, I am slightly inclined to blame the user :). Even searching for "No valid files found in content" gets you the relevant information in the first hit... But, will extending warning a bit more be helpful? Something like:
> pelican content
WARNING: No valid files found in content
| Active readers:
| * BaseReader (static)
| * HTMLReader (htm, html)
| * RstReader (rst)
|
| Inactive readers (explicitly deactivated or missing dependency):
| * MarkdownReader (md, markdown, mkd, mdown)
Done: Processed 0 articles, 0 drafts, 0 pages, 0 hidden pages and 0 draft pages in 0.07 seconds.
Yes, that looks good. I don't think this should be a high priority, just something that could be a bit nicer.
And, I know the temptation (re users) Most of the time I like it when there are no (other) users :-) Thanks for all the work on Pelican!
I had the same issue as terrycojones. Migrated to a new machine, installed pelican, but no markdown (yet). Did a make command to generate AND publish. Poof, website content was all gone... Should have been a bit more careful. Luckily I found this thread within a minute. Thumbs up for explicitly stating markdown is not installed. Maybe especially if (like in my case) markdown files are found in the content dir.
Even though the documentation is extremely clear on that?
Is is not extremely clear, the quickstart guide mentions nothing about it.
One idea could be to have a checkhealth command that checks for this kind of "issue".
@kaddkaka Which quickstart are you looking at? It's mentioned right at the start:
Install Pelican (and optionally Markdown if you intend to use it) ...
For what it's worth, I did not find the error message particularly explicit (something mentioning markdown and/or pip install would have helped).
WARNING No valid files found in content for the active log.py:91 readers: | BaseReader (static) | HTMLReader (htm, html) | RstReader (rst)
I'm just trying pelican, I didn't go deeply in the docs yet and just read a short guide with commands, I haven't wasted 30 minutes because I found this issue in 30 secs but I agree with @JBacc1 adding something "maybe you need markdown installed if you are using .md files" in the warning would not be a big issue and help
