Skip to content

core Module Docs

Jump to bottom Edit New page
cjlee112 edited this page Sep 1, 2013 · 1 revision

The spnet data model

This module defines the spnet data collections. These are stored as MongoDB collections:

  • core.Person: a participant in the spnet, either because they've signed in to selectedpapers.net, or wrote recommendations / posts etc.
  • core.Paper: a paper, either from arXiv, Pubmed, or the DOI database. Note that since the primary activity on selectedpapers.net is working with a paper and its associated data, we embed these associated datatypes (e.g. Recommendation) in the Papers collection.
  • core.SIG: a Specific Interest Group, i.e. a research topic.
  • core.GplusSubscriptions: lists a Google+ member's subscriptions (to other Google+ members).

The following data are embedded in Paper documents in MongoDB:

  • core.ArxivPaperData: information about an arXiv paper.
  • core.PubmedPaperData: information about a Pubmed paper.
  • core.DoiPaperData: information about a DOI paper.
  • core.Recommendation: an spnet recommendation for a paper, typically tagged for one or more topics.
  • core.Post: a post discussing a paper, typically tagged for one or more topics.
  • core.Reply: a reply to a post.
  • core.PaperInterest: tags a paper as interesting for a specific person, in one or more topics.

The following data are embedded in Person documents in MongoDB:

  • core.GplusPersonData: Google+ information about a specific Person.
  • core.EmailAddress: email address associated with a Person.
  • core.Subscription: a subscription for a given person, to receive recommendations from a specific person, typically for one or more topics.

Paper Data

  • Paper(fetchID=None, docData=None, insertNew=True)

    Subclass of base.Document; each instance represents a paper. Guarantees the existing of the following attributes:

  • Paper.title: Title of the paper.

  • Paper.authorNames: A list of strings each representing an author name.

    In addition, regardless of the source of the paper, you can call the paper's get_value() method with any of the following argument values:

    • 'spnet_url': full selectedpapers.net URL for the current paper, e.g. https://selectedpapers.net/arxiv/1234.5678
    • 'local_url': path for the current paper, e.g. /arxiv/1234.5678
    • 'source_url': URL for the source database's page for this paper.
    • 'downloader_url': URL for full-text access (if available).
    • 'doctag': ID tag for this paper
    • 'abstract': abstract for this paper, if available.

Typically also contains one or more of the following subdocuments:

  • Paper.arxiv: core.ArxivPaperData object representing arXiv record for this paper from arXiv.

  • Paper.pubmed: core.PubmedPaperData object representing Pubmed record for this paper from Pubmed.

  • Paper.doi: core.DoiPaperData object representing DOI record for this paper.

  • Paper.recommendations: list of core.Recommendation objects for this paper (if any).

  • Paper.posts: list of core.Post objects for this paper (if any).

  • Paper.replies: list of core.Reply objects for this paper (if any).

  • Paper.interests: list of core.PaperInterest objects for this paper (if any).

  • Paper.get_interests(people=None, sorted=False): Returns dictionary of topics for which people have tagged this paper as interesting. The keys are core.SIG topic objects, and their associated values are lists of people (core.Person objects) who tagged the paper as interesting for that topic.

    people, if not None, must be a set of people for filtering the results; i.e. only topics tagged by those people will be returned.

    if sorted is True, the results will be returned as a list of (topic, people) tuples sorted in descending order of len(people).

Person data

  • Person(fetchID=None, docData=None, insertNew=True)

    Subclass of base.Document; each instance represents a person. Guarantees the existence of the following attributes:

  • Person.name: string representing the person's name.

    In addition, regardless of the source of the person record, you can call its get_value() method with any of the following argument values:

    • 'spnet_url': full selectedpapers.net URL for this person, e.g. https://selectedpapers.net/people/123456
    • 'local_url': path for this person, e.g. /people/123456

Typically also contains one or more of the following subdocuments:

  • Person.gplus: core.GplusPersonData object representing Google+ record for this person.

  • Person.subscriptions: list of core.Subscription objects representing this person's current subscriptions.

    Also provides attributes that link to other documents ("foreign keys"):

  • Person.papers: Papers for which this person is an author, as a list of core.Paper objects.

  • Person.recommendations: Recommendations written by this person, as a list of core.Recommendation objects.

  • Person.subscribers: People who have subscribed to this person, as a list of core.Person objects.

  • Person.posts: Posts written by this person, as a list of core.Post objects.

  • Person.replies: Replies written by this person, as a list of core.Reply objects.

  • Person.interests: Papers tagged as interesting by this person, as a list of core.PaperInterest objects.

  • Person.readingList: Papers added by this person to their reading list, as a list of core.Paper objects.

  • Person.get_interests(sorted=False): Returns dictionary of topics for which this person tagged papers as interesting. The keys are core.SIG topic objects, and their associated values are lists of papers (core.Paper objects) tagged as interesting for that topic.

    if sorted is True, the results will be returned as a list of (topic, papers) tuples sorted in descending order of len(papers).

  • Person.update_subscribers(klass, docData, subscriptionID): To be called when new Person first added to the database. Converts other peoples' pending subscriptions to this person, to a new entry in their subscriptions field.

Google+ Person data

  • GplusPersonData(fetchID=None, docData=None, parent=None, insertNew=True): Subclass of base.EmbeddedDocument; each instance stores Google+ data for one person. Detailed information about its content is available from Google. Here we just list some key fields:

  • GplusPersonData.id: String representing the person's Google+ unique ID.

  • GplusPersonData.displayName: String representing name to display for this person.

  • GplusPersonData.update_posts(maxDays=20, **kwargs): get new posts from this person, updating old posts with new replies. Queries Google+ for the data (for changes up to maxDays old), and saves any changes (as Recommendations, Posts, Replies) to the database.

  • GplusPersonData.update_subscriptions(doc, subs): Update the core.GplusSubscriptions record for this person with new data subs, and then use them to add to our Person.subscriptions.

    doc must be the response document from Google+ People list request; subs should be iterator object for the corresponding list of people.

Topic data

  • SIG(fetchID=None, docData=None, insertNew=True): Subclass of base.Document; each instance represents a topic. Guarantees the existing of the following attributes:
  • SIG.name: string representing the topic name. This is just its _id with a hash (#) prefix. I.e. if _id is numberTheory, then name is #numberTheory.
  • get_value(): you can call this method with any of the following argument values:
    • 'spnet_url': full selectedpapers.net URL for this topic, e.g. https://selectedpapers.net/topics/numberTheory
    • 'local_url': path for this topic, e.g. /topics/numberTheory

Also provides attributes that link to other documents ("foreign keys"):

  • SIG.recommendations: Recommendations tagged for this topic, as a list of core.Recommendation objects.
  • SIG.posts: Posts tagged for this topic, as a list of core.Post objects.
  • SIG.interests: Papers tagged as interesting for this topic, as a list of core.PaperInterest objects.
  • SIG.get_interests(): Returns dictionary of papers tagged as interesting for this topic. The keys are core.Paper objects, and their associated values are lists of people (core.Person objects) who tagged the paper as interesting for this topic.

class methods

  • SIG.standardize_id(fetchID): Checks whether fetchID fits hashtag character rules. Returns the _id value for the new topic (i.e. with # character stripped off), or raises KeyError if it is invalid as a hashtag.
  • SIG.find_or_insert(fetchID, published=None, **kwargs) save new topic to db if not already present, after checking its hashtag format validity.
Add a custom footer
Add a custom sidebar
Clone this wiki locally