Giters

S2-group / NAPPA

Implementation of a navigation-aware technique for personalized prefetching of network requests of Android apps

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Refactor the Room classes organization

github-actions opened this issue · comments

commented

Refactor the Room classes organization

The current structure is messy and falling apart. The Dao/Entity/View model could be

replaced in favor of the domain/feature model as done for the activityVisitTime feature.

This organization allows to keep together classes that are related to each other, making

maintenance simpler.

While we are at it, consider renaming:

  • The class PrefetchingDatabase to NappaDB;

  • The class SessionData to ActivityVisitFrequency or simply VisitFrequency;

  • The entity classes MyEntityData to just MyEntity;

  • The database schema pf_* to nappa_*;

  • The database file pf_db to nappa_db.db --> Verify if it is possible to add the extension;

Other considerations:

  • Enforcing required data by either making entities attributes type primitive or annotating object types with @nonnull;

  • Enforcing foreign keys at the database level --> currently only enforced in the application levels. Adding invalid FK is currently allowed;

  • Extracting model class UrlCandidateToUrlParameter from UrlCandidateDao;

  • Extracting model class SessionAggregate from SessionDao;

  • Using DatabaseView to simplify some queries;

  • Extracting LAR operations from ActivityTableDao --> LAR refers to PageRank, HITS and Salsa scores, but I am not sure what the acronym means;

  • Renaming the method getInstance(Context) to init(Context) since its purpose is to initialize the DB and it must know the Context;

  • Resolve warnings listed when building the library;

The model classes extraction is motivated to simplify code reading. It is a lot better to read

List<SessionAggregate> than List<SessionDao.SessionAggregate> .

There are probably other improvements and considerations to take into account, but this is the

preliminary list after working in the NAPPA library for a few weeks.

If not familiar with the database, I recommend using an app with NAPPA enabled for a while

to create data and then open the database in a SQL tool (e.g., DB Browser for SQLite).

NAPPA/Prefetching-Library/android_prefetching_lib/src/main/java/nl/vu/cs/s2group/nappa/room/PrefetchingDatabase.java

Line 26 in 64c21d0

// TODO Refactor the Room classes organization 

import nl.vu.cs.s2group.nappa.room.converter.DateConverters;
import nl.vu.cs.s2group.nappa.room.dao.ActivityExtraDao;
import nl.vu.cs.s2group.nappa.room.activity.visittime.ActivityVisitTimeDao;
import nl.vu.cs.s2group.nappa.room.dao.GraphEdgeDao;
import nl.vu.cs.s2group.nappa.room.dao.SessionDao;
import nl.vu.cs.s2group.nappa.room.dao.UrlCandidateDao;
import nl.vu.cs.s2group.nappa.room.data.ActivityExtraData;
import nl.vu.cs.s2group.nappa.room.activity.visittime.ActivityVisitTime;
import nl.vu.cs.s2group.nappa.room.data.LARData;
import nl.vu.cs.s2group.nappa.room.data.Session;
import nl.vu.cs.s2group.nappa.room.data.SessionData;
import nl.vu.cs.s2group.nappa.room.data.UrlCandidate;
import nl.vu.cs.s2group.nappa.room.data.UrlCandidateParts;
import nl.vu.cs.s2group.nappa.room.activity.visittime.AggregateVisitTimeBySession;

// TODO Refactor the Room classes organization
//  The current structure is messy and falling apart. The Dao/Entity/View model could be
//  replaced in favor of the domain/feature model as done for the `activityVisitTime` feature.
//  This organization allows to keep together classes that are related to each other, making
//  maintenance simpler.
//  While we are at it, consider renaming:
//  * The class `PrefetchingDatabase` to `NappaDB`;
//  * The class `SessionData` to `ActivityVisitFrequency` or simply `VisitFrequency`;
//  * The entity classes `MyEntityData` to just `MyEntity`;
//  * The database schema `pf_*` to `nappa_*`;
//  * The database file `pf_db` to `nappa_db.db` --> Verify if it is possible to add the extension;
//  Other considerations:
//  * Enforcing required data by either making entities attributes type primitive or annotating object types with @NonNull;
//  * Enforcing foreign keys at the database level --> currently only enforced in the application levels. Adding invalid FK is currently allowed;
//  * Extracting model class `UrlCandidateToUrlParameter` from `UrlCandidateDao`;
//  * Extracting model class `SessionAggregate` from `SessionDao`;
//  * Using DatabaseView to simplify some queries;
//  * Extracting `LAR` operations from `ActivityTableDao` --> LAR refers to PageRank, HITS and Salsa scores, but I am not sure what the acronym means;
//  * Renaming the method `getInstance(Context)` to `init(Context)` since its purpose is to initialize the DB and it must know the Context;
//  * Resolve warnings listed when building the library;
//  The model classes extraction is motivated to simplify code reading. It is a lot better to read
//  `List<SessionAggregate>` than `List<SessionDao.SessionAggregate>` .
//  There are probably other improvements and considerations to take into account, but this is the
//  preliminary list after working in the NAPPA library for a few weeks.
//  If not familiar with the database, I recommend using an app with NAPPA enabled for a while
//  to create data and then open the database in a SQL tool (e.g., DB Browser for SQLite).
@Database(
        entities = {
                RequestData.class,

b7497ee39326b0a622cbbfb8f6e64f39e8f5bd70

commented

Replaced by issue #97