jamescheney / database-wiki Goto Github PK

Debug/review and merge the chart-building and mapping features from Haoli's 
project into the main distribution.

Review/debug and merge provenance querying features from Snehal's project into 
main distribution.

Add support for exporting entries as JSON.

Step 1.  Merge trunk changes into a copy of Hui's project branch.  See if this 
works.

Step 2.  Implement simple way of re-indexing when an entry changes.

Step 3 (future?) Implement smarter re-indexing.

Extend the import to allow merging of external documents or previously exported 
wiki snapshots into a wiki.

Logging in doesn't work properly:

1.  If you log in with an invalid user id, there is no failure message, instead 
from then on the login link is a no-op/page forward.

2.  If you log in with a valid user id but incorrect password, it succeeds.  
(should fail/give an error message).

3.  There is no way to log out short of restarting the browser.

The template mechanism should provide a symbol @imports so that we can add new 
javascript library imports in one place (in the code) instead of once per 
template.

Otherwise we will have to change all the templates manually whenever the 
javascript code we're importing changes.

The default and built-in templates should be changed to take this into account 
too.

What steps will reproduce the problem?
1.
2.
3.

What is the expected output? What do you see instead?


What version of the product are you using? On what operating system?


Please provide any additional information below.

Develop a binary (JAR) distribution and installation script that puts the 
program somewhere global and runs the server from there rather than running 
from the source directory.

Allow schema evolution and track changes to schema.

Attempting to insert schema elements in the App Engine version leads to failure 
with the error: "There are no generated keys."

Seems to be due to getGeneratedKeys not returning anything, so maybe something 
isn't defined as auto_increment in Cloud SQL that should be.

Add buttons or links that generate XML, JSON (or eventually other) views of the 
data.

Currently, wiki pages are lost when we refresh a database and there is no 
straightforward way to export them and re-import them later.  

Likewise, the (new) config file for URL generation is not being 
imported/exported.

Loading a large dataset through the web interface on the App Engine version 
leads to an error (the connection is dropped before the data is finished 
loading).

App Engine limits the time per request to 60 seconds.  

To do: investigate use of backends or tasks to handle imports that take longer 
than 60 seconds.

Use path-based URLs instead of unique identifiers to locate data

What steps will reproduce the problem?
1. Search for an unbalanced string, such as "\""foo".
2. Wait.

Server never responds.

When deletion is forbidden by the access control policy, an "access denied" 
message is shown but the deletion takes place anyway.

Currently, wiki pages are associated with collections and can only query data 
in that collection.  

This enhancement would make the "wiki page" component of the system a 
server-level object rather than having several Wiki objects, one per 
collection.  It would also allow writing queries that access data from more 
than one collection (which might in turn require some modifications to the 
schema language to provide top-level variable bindings for the different 
collections).

This would also mean that we could write documentation for the whole system as 
(global) wiki pages, making it more self-documenting.

Data import is currently broken using the schema archiving  For example, using 
attached trivial file foo.xml, I get:


$ sh database-import.sh FOO FOO foo.xml jcheney
[TIME] [Duplicate version name]: Version 8 May 2011 17:06:49 already exists
    at org.dbwiki.data.time.version.VectorVersionIndex.add(VectorVersionIndex.java:70)
    at org.dbwiki.driver.rdbms.DatabaseImportHandler.endImport(DatabaseImportHandler.java:64)
    at org.dbwiki.data.io.XMLDocumentImportReader.start(XMLDocumentImportReader.java:106)
    at org.dbwiki.web.server.WikiServer.registerDatabase(WikiServer.java:717)
    at org.dbwiki.main.DatabaseImport.main(DatabaseImport.java:91)

This also happens if I try to add data through the web interface.

The authorization model in the "trunk" version (just merged in) doesn't seem to 
pay attention to the fine-grained authorizations (per user/database/entry), 
only the top-level access flags governing the whole dataset.  

The schema language isn't documented, and it's hard to find out what the schema 
of an existing database is.

Allow import of layout information from other wikis.

To reproduce:

Click "New"

Enter "TEST" and "Test" for short name/name

Enter a url to some XML

Click "Save"

This will lead to a page with a red box around the schema and message:

Please specify the target pat and schema for the given resource. 

This isn't very helpful; it would be mroe helpful if it said something like: 
"You didn't provide a schema, so we have inferred one that seems to work for 
your document.  Click "Save" if this is OK, or edit the schema."

Document the underlying database schema used in the trunk version, so that 
students can understand it and do projects.

1. Add capability to export entries (or parts of entries) as RDF.  

2. Add capability to export provenance, annotation, history information as RDF, 
cross-linked against the data.