Cloud Firestore

Backend of [Regroup

• The old KJ method tool used Google Realtime API, but will end in January 2019

• I was thinking about using Firebase, but when I investigated, Cloud Firestore had a beta release.

  • I decided to go this way because it seemed to fit better in terms of the data model.

• I’ve dropped the idea of using Scrapbox as a backend.

  • Scrapbox is still an interface to humans, not a place to put mechanical data.
    • Import from Scrapbox will be made available.

• How “Documents” are in “Collections

  • Think of the document as JSON (+ some additional data types).
  • Documents up to 1 MB
  • I can put the collection back in the document.

• supported data types | Firebase

  • Maps, arrays, and references can be included in the document.
  • text

    • Only the first 1,500 bytes of the UTF-8 representation are considered by the query.

    • Indexes on single fields are automatically generated, but long texts should be explicitly excluded (in terms of billing) because indexes are meaningless

• doc(”…”) .set(…) Overwrite with (or create if not present)

  • merge: merge if true is specified ref
  • collectionsRef.doc(“hoge”).set(…) can be created by specifying an ID with
  • collectionRef.add(…) can be added with automatic ID generation
    • ref = collections.doc() and then ref.set(data) would be better
    • {timestamp: firebase.firestore.FieldValue.serverTimestamp()} would be good to include
      • I may write offline, so I’d better put the timestamp at hand in the CREATED field.
  • Write instruction returns Promise and then is called on completion
    • If you write offline, then when you come back online and can write to the server, then is called.
    • So it seems like a good idea to watch this and show the “written” mark
    • Apart from that, snapshot.metadata.hasPendingWrites will tell you if there are any outstanding writes.
      • Get real-time updates with Cloud Firestore | Firebase
  • There’s also docRef.update, what’s the difference between that and merge:true?

• Getting a document that does not exist does not result in an error.

  • Empty objects return.
  • doc.exists() ⇒ false

• .onSnapshot(function(querySnapshot) {

  • Listen for updates

• Batch Write src

  • 500 at a time
  • batch = operation against db.batch() then batch.commit()

• full text search | Firebase

• Offline persistence

  • Enable Offline Data | Firebase

  • If a user has multiple browser tabs open that reference the same Cloud Firestore database and offline persistence is enabled, Cloud Firestore will only work correctly on the first tab.

  • I’m afraid that would be an inadvertent mistake, opening multiple tabs and skimming the data.
  • The sample code seems to be able to detect it at least, so we should warn users.
  • I’ve confirmed that what I write off wifi is saved on the server after reconnecting wifi.
    • Whether there is a connection or not, snapshot.metadata is fromCache: false
    • On the other hand, when the browser is reloaded, fromCache: true

• Python API

  • Try using Cloud Firestore | Firebase in Python.
  • For cred = credentials.ApplicationDefault() to succeed Start Authentication | Authentication | Google Cloud to download JSON and set environment variables.
  • In Python, collection.document() instead of collection.doc().
  • DocRef#get returns `document.DocumentSnapshot
    • .to_dict() to become a dictionary
  • It says that db.collection("...") .get() says “returns all documents” and I thought it was serious, but it creates a generator, so no worries!
    • DeprecationWarning: 'Collection.get' is deprecated: please use 'Collection.stream' instead.
    • xs.__next__() # => document.DocumentSnapshot
  • Use where for conditions python

In [25]: [d.to_dict()["text"] for d in db.collection("pieces").where("text", "<", "t").stream()]
Out[25]: []

In [26]: [d.to_dict()["text"] for d in db.collection("pieces").where("text", ">", "t").stream()]
Out[26]: ['test']

- [https://googleapis.github.io/google-cloud-python/latest/firestore/client.html](https://googleapis.github.io/google-cloud-python/latest/firestore/client.html)

• Web JS API
  • Add Firebase to your JavaScript project | Firebase
    • First read some JS in the script tag, then configure
    • They have a JS that won’t work if you open it from the local file system.
      • Serve with firebase-tools
      • I’m sure a Python http.server would be fine, but first I decided to do as you said.

      • Error: Cannot understand what targets to deploy. Check that you specified valid targets if you used the —only or —except flag. Otherwise, check your firebase.json to ensure that your project is initialized for the desired features.

      • hmm
      • $ python3 -m http.server
        • It doesn’t seem particularly problematic in
      • It seems easier to use Firebase Hosting | Firebase at release time
  • Using Cloud Firestore | Firebase
    • In JS, set and get return Promise js

db.collection("pieces").get().then((querySnapshot) => {
    querySnapshot.forEach((doc) => {
        console.log(`${doc.id} => ${doc.data()["text"]}`);
    });
});

• REST API
  • https://firebase.google.com/docs/firestore/reference/rest/v1beta1/projects.databases.documents/createDocument

---

This page is auto-translated from [/nishio/Cloud Firestore](https://scrapbox.io/nishio/Cloud Firestore) using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.