bobinette / papernet Goto Github PK

on clean setup....
seems i need a config.

x-MacBook-Pro:papernet apple$ mkdir data
x-MacBook-Pro:papernet apple$ go run cmd/cli/*.go index create --index=data/papernet.index --mapping=bleve/mapping.json
Index created at data/papernet.index
x-MacBook-Pro:papernet apple$ go run cmd/web/main.go
../../gin-gonic/gin/binding/default_validator.go:7:2: cannot find package "gopkg.in/go-playground/validator.v8" in any of:
	/usr/local/opt/go/libexec/src/gopkg.in/go-playground/validator.v8 (from $GOROOT)
	/Users/apple/workspace/go/src/gopkg.in/go-playground/validator.v8 (from $GOPATH)
x-MacBook-Pro:papernet apple$ go get gopkg.in/go-playground/validator.v8
x-MacBook-Pro:papernet apple$ go run cmd/web/main.go
FATA[0000] could not open key file: open configuration/hs256.json: no such file or directory  env=dev
exit status 1
x-MacBook-Pro:papernet apple$ go run cmd/cli/*.go index create --index=data/papernet.index --mapping=bleve/mapping.json
Index data/papernet.index already exists, no need to create it
x-MacBook-Pro:papernet apple$ go run cmd/web/main.go
FATA[0000] could not open key file: open configuration/hs256.json: no such file or directory  env=dev
exit status 1

so i figured maybe the CLI allows me to create my auth key. But no luck..

x-MacBook-Pro:papernet apple$ go run cmd/cli/*.go auth
FATA[0000] could not open key file: open configuration/hs256.json: no such file or directory  env=dev
exit status 1

this is nice for tracking papers.

but am lost with setting up locally.

can you add basic step to setup in readme or here ?

When trying to sign up I get the following error:
Could not exchange token: Unexpected end of JSON input

Check if go-kit could be used for Papernet, especially this example

The idea would be to filter by tags, but by exact match, i.e. that if I have two papers:

• Machine learning basics - tags: Machine learning
• Learning how to build your time travelling machine - tags: Mechanics, Sci Fi

Then the second would not match a search with "machine learning".

Apart from helping fine tuning the search, it could help creating topics or collections, e.g. "Machine learning for dummies" or "To read".

This feature would require a new field in the index that is not analysed, and to add a facet on this field. The list on tags already works on exact match like that, so there should be no need to rework it.

More specifically, in the case of books, it could be nice to have a "paper" for each chapter. The main advantage would be the exercices

When I search for a paper on the arxiv tab (search for A new quantum ripple-carry addition circuit), I find the paper but when I click on see on papernet, I just end up on the page of one of my other papers, without having the paper added to my space

is Master meant to be stable ?
I got many errors.

Hi,

Hope you are all well !

Would be awesome to have a docker-compose file to bootstrap the front-end,
https://github.com/bobinette/papernet-front, and the back-end. And if they are Alpine based to have small containers it would be really awesome. I can do that but I would need a more simple auth mode for test/demo purpose.

Ps. Have you seen these references ?

• https://github.com/feedlabs/elasticfeed
• https://github.com/feedlabs/feedify
• https://github.com/urandom/readeef

they could help you to aggregate more sources and have a better semantic search mode.

Have a great day !

Cheers,
Richard

Needs:

• Owner on paper
• Teams in db:

type Team struct {
    ID int
    Members []int
    Admins []int
}

• Admins can add new members to a team
• Give access to a paper to the members of a team, base on canEdit

Hello Bobinette,

I love your Papernet project. Great idea !

I encountered a minor (but somehow annoying) issue when editing a paper.
Blinking cursor is frequently sent to bottom of the page in edition mode (it may take a few modifications).
For that reason, when editing a paper, I can see the text above the cursor, but not what is after the cursor (unless I scroll down again).
Sometimes it works properly. See gif.

Have a great day !

Cheers,

Fraxœ

When I search in the list or the select for papers, the search uses whole words. It would much more intuitive to use starts with or fuzzy match.

Have a look at:
https://medium.com/wireless-registry-engineering/short-introduction-to-bleve-5de4bbf16657#.o2k85xq3i

I want an API that takes an URL as input and returns a paper as JSON, which I can then use directly to create a new paper through the paper creation API

Currently the access_denied error from google auth is not handled.

Currently a new paper is created with some default values, ignoring any JSON that could be sent.

Now that I need the email, I need to ask the user to log out and log in again to get it from Google. It would be better to be able to refresh the user information through a Papernet end point.

Check this link.

I can only state whether I have read or not a document. It works quite well for papers, but is not very accurate for books. Adding a "Reading" status would be nice

Use the different arguments that arxiv's API exposes, instead of using the all argument for every query.

Do you know about storm ?

It's pretty nice way to wrap access to boltdb. It's fairly well used.

Currently, the papers are sorted by _id in the bleve query. The problem is, because the doc ID field is a string the doc 16 ends up before the document 2.

x-MacBook-Pro:papernet apple$ ./cli paper all
2017/05/21 11:45:46 error opening db: timeout

the addresss is using is "data/papernet.db"

Also i cant go run main.go in the cli folder, because of the way you setup the cobra cmd stack.
Its not a killer, but it does make it a bit tougher to iterate.
Might be worth setting up a makeFile in the root that runs the standard developer things.
like building the CLi and copy it to the root directory so its can find the DB path etc.