yymao / kipac-teabot Goto Github PK

When trying to find somebody to read 1504.01734, I get the following error:

Error: 'ascii' codec can't encode character u'\u03b1' in position 75: ordinal not in range(128)

See http://stanford.edu/~yymao/cgi-bin/kipac-teabot/find-someone#1504.01734 .

How about a link at the bottom of the teabot mail, that gives the full day's arxiv postings sorted by teabot relevance to me?

Hi, I have 3 arXiv "names" and would like to add all 3 to the TeaBot. For now I just filled out the form 3 times. Is there a better way to do this? Thanks!

prune the list of KIPACs to only those who are active and local

Let subscribers add or remove particular papers that are used in their topic model through a web interface.

It would be nice to have two options: interesting and not interesting.
It may be useful to collect the response before and after reading the abstract / paper (to understand whether people change their mind after reading the paper).

I think this is open to discussion, not sure if there is one optimal way to do it, but seems it has to be decided for now to get the tool rolling -- do we use the intersection of the top several in each to create the set of 3 that the Teabot will mail, or choose the top one of each, or some other way..?

In principle, there should be a date period associate with each KIPAC member, we can then take care of short-term visitors. We still need an updated member list that includes short-term visitors and the dates.

When an user clicks "want to hear/talk about a paper" on the response page, TeaBot should inform tea organizers by sending an email to tealeaks.

The rationale is:

• Cc'ing several people might encourage email discussion of the paper
• One paper per email enables us to put the paper title in the email subject line
• To assess whether I am interested in a paper, I just want to read th eabstract without clicking away from my email
• Clicking on the link registers a vote of interest with TeaBot: I want people to vote for papers by clicking through to the PDF, not the arxiv abstract, because I want to know who has actually read the paper! (We can assume opened PDF = "read" for tea purposes ;-)

Thoughts?

When a case of author misattribution is reported, the fix should be patched into the topic model and discovery database.

All KIPAC members should be able to opt in (either by sending an email to Yao or sign up on a web interface) for TeaBot's daily paper suggestion service.

Martha can provide lists of people on demand!

Hey @yymao - the KIPAC TeaBot is extremely useful code, and other institutes are showing interest in running it. Can you comment briefly on how hard it would be to prepare it for general release, and how much maintenence you think that would take? Also, you'll need someone at KIPAC to maintain the system after you've gone - could be good to involve that person in the work (documentation, testing etc) leading up to release. One way to absolve yourself of responsibility coudl be to transfer ownership to KIPAC, where the repo would be KIPAC/TeaBot. However, this would diminish your visibility as its creator! I guess we'd just write your name and website all over it :-)

Personal preferences such as email frequency, and number of papers by email.

To include people who usually don't show up in Tea but still affiliated with KIPAC.

This is primarily for the blog team to choose a paper to highlight - but also for advertising on the KIPAC webpage. We'll figure out embedding frames later - main thing is to get some kind of web page up!
NB we need to track papers by staff and faculty as well, so we'll need these lists from Martha too.

Unfortunately the PDF linked to from the Readmore button does not work in my iPhone, which too often is the place I see my email on. Is there an easy fix?

eg:

Here's a list of new arXiv papers authored by KIPAC members:

Eli Rykoff: [1501.02840] Environment-based selection effects of Planck clusters

Many congratulations, Eli!

This message is automatically generated and sent by KIPAC TeaBot.
Create an issue if you have any suggestions/questions.

When one clicks on the like/dislike buttons on the response page, TeaBot should integrate these responses into the topic model of that person.