Google AppEngine

The mapper API isn't just limited to mapping over datastore entities, either. You can map over lines in a text file in the blobstore, or over the contents of a zip file in the blobstore. It's even possible to write your own data sources

I was very pleased to see that the new OAuth API provides the same interface as the Users API, but I see a missed opportunity here:

my problem is that there are two versions of the method: google.appengine.api.oauth.get_current_user() google.appengine.api.users.get_current_user() one works with OAuth and one with sign-in (or am I wrong?).

Well java.lang.System is whitelisted, so won't java.lang.System.out.println(...) work?

The property map consists of a set of 'property' entries, each of which specifies how to handle a particular property of the model on import and on export. for our Permission kind, the bulkloader has identified 4 properties, plus the __key__ pseudo-property. Each has an 'external_name', and optional import and export transforms, which specify how to translate between the App Engine datastore representation and an external representation.

All we had to do here was to remove some of the boilerplate and the extraneous invite_nonce entry, and fill in the kind names for the two reference properties, and we're sorted.

we didn't have to write a single line of Python code, or set up an app.yaml, or anything else Python-specific in order to achieve it! Further, the bulkloader took care of generating a mostly-finished configuration file for us, in a format that ensures the data we download can be re-uploaded again without loss of fidelity.

Yes, you will need to use the federated login (OpenID) stuff. The long-and-short of it is that you pass the federated_identity parameter to users.create_login_url.  You'll need to setup a page for users to tell you what goes in federated_identity somehow, perhaps by clicking a google logo or entering an apps domain. For Google accounts:   users.create_login_url(federated_identity='google.com/accounts/o8/id')   or   users.create_login_url(federated_identity='gmail.com') For an Apps account:   users.create_login_url(federated_identity='google.com/accounts/o8/site-xrds?hd=yourappsdomain.com') There is a little info here:   http://code.google.com/appengine/docs/python/users/overview.html And Wesley has a nice article about it here:   http://code.google.com/appengine/articles/openid.html Some info on Google Apps domains and OpenID:   http://groups.google.com/group/google-federated-login-api/web/openid-...

Because A is the entity group parent of B, the key of an instance of B must contain information about its entity group parent instance of A. Instead of using b.key.getId() in   // We cannot retrieve this object. Why?   B newB = pm.getObjectById(B.class,b.key.getId()); you might care to use the KeyFactory.Builder class as detailed in "http://code.google.com/intl/en/appengine/docs/java/datastore/ creatinggettinganddeletingdata.html#Creating_and_Using_Keys" in order to build a key instance which contains information about both your B instance and its parental A instance.

Here is the approach I would probably take unless you have many hundreds of child tasks. I would keep a list of keys of all children of each project/task. (You can then do a db.get and get all immediate children with single call) and each child has a reference to it's parent. In addition in each item keep a list of all parents keys back to the root. That way you can search for all children in part of the tree/subtree (This approach is especially useful if you have multiple child types) as back ref sets are kind specific. Keep a  running count of complete children at each level, rather than trying to calculate the current state in a single big function,  (keep these subsidiary values up to date through tasks)

The next problem that arises is that if the database has grown too big, then the csv generation will be slow(as it has to jump a lot of times). So the next idea is , instead of generating all cursors in the main cgi , the main cgi generates only first few set of page links, and then triggers a task queue entry to continue from there. By the time the user would have read the first page and clicks on "next page", the task would have generated few more entries. As the set of cursors for pagination is in memcache, you can use it even in your first page, if you are using AJAX (guessing, not that sure.)

If using the datastore API directly, a common pattern of usage is: // Get a handle on the datastore itself DatastoreService datastore = DatastoreServiceFactory.getDatastoreService(); // Lookup data by known key name Entity userEntity = datastore.get(KeyFactory.createKey("UserInfo", email)); // Or perform a query Query query = new Query("Task", userEntity); query.addFilter("dueDate", Query.FilterOperator.LESS_THAN, today); for (Entity taskEntity : datastore.prepare(query).asIterable()) { if ("done".equals(taskEntity.getProperty("status"))) { datastore.delete(taskEntity); } else { taskEntity.setProperty("status", "overdue"); datastore.put(taskEntity); } }

it could be implemented as follows: The developer creates a CName record from *.<developersAppSubDomain>.<developersDomain>.com to <appEngineAppId>.appspot.com (if wildcard CName values aren't available on the developer's DNS server, then the developer will have to create one entry per customer). When Google receives a request on the App Engine Server with an unknown host name such as <subdomain>.<customersDomain>.com, it does the CName look-up, until it finds a host name of the form <appEngineAppId>.appspot.com. It then sends the server request on to that App Engine application, ideally including the whole CName "chain" as a special CGI variable, so that the application can determine the user ID of the customer that owns the domain. (And if the App Engine server doesn't find a host name for an App Engine application, it returns 404.)

the necessary changes to appengine-web.xml to get Sitebricks to run in development mode:         <sessions-enabled>true</sessions-enabled>         <system-properties>                 <property name="java.util.logging.config.file" value="WEB-INF/logging.properties"/>                 <property name="mvel2.disable.jit" value="true"/>         </system-properties>

the necessary changes to appengine-web.xml to get Sitebricks to run in development mode:         <sessions-enabled>true</sessions-enabled>         <system-properties>                 <property name="java.util.logging.config.file" value="WEB-INF/logging.properties"/>                 <property name="mvel2.disable.jit" value="true"/>         </system-properties>

We had a GaeFlashCache built specifically for appengine. Try doing: bind(FlashCache.class).to(GaeFlashCache.class).in(Singleton.class);

the necessary changes to appengine-web.xml to get Sitebricks to run in development mode:         <sessions-enabled>true</sessions-enabled>         <system-properties>                 <property name="java.util.logging.config.file" value="WEB-INF/logging.properties"/>                 <property name="mvel2.disable.jit" value="true"/>         </system-properties>

As I understand it, the process of performing a single fetch (call to get())  from the dastastore using a key basically involves finding the host housing the entity, opening a socket, fetching the data, and then cleaning up the connection.  So to fetch something like 30 entities from the datastore, you're repeating the process 30 times over in serial, each time incurring whatever overhead is involved.  I also read that if you perform bulk fetches, (ie passing multiple keys at once) you can eliminate a great deal of that overhead.  In one of the videos I watched from Google I/0 2009, the presenter (whose name I forget - d'oh) said that performing a bulk fetch actually performs the fetches in parallel from the data store and you shoudl see requests noticeably faster.

I figure it is going to take me at least four iterations to get this right. The first one will be building a GWT application with a simple UI that has no server logic behind it (just to learn how layout in GWT works). Step two will be adding a fake servlet backend (not app engine, just in memory). While not exactly App Engine yet, I should have a completely specified client-server API by the end of this process that I can subsequently implement on App Engine (iteration 3). Iteration four will handle deployment, CSS and whatever I may screw up in iterations one and two. I will log my notes of things I run into while I code.

In a relational world duplication is removed in order to prevent update anomalies. Error prevention is the driving force in relational modeling. Normalization is a kind of ethical system for data.

BigTable data ethics are more Mardi Gras than dinner with the in-laws. Data just wants to have fun. BigTable won’t stop you from hurting yourself. And to get the best results you may have to engage in some conventionally risky behaviors.

list properties are indexed as 'multiply valued properties', with one index row per list entry. The index you specify should work fine.

You can tell if it's using sqlite by executing 'sqlite3 /path/to/datastore' on the command line, and seeing if sqlite recognizes it as a valid DB. The code is essentially the same as that which I announced on my blog, so there shouldn't be significant performance differences.

'm guessing your tests were run locally because the counter in the local datastore does indeed have datastore scope.  The scope of the counter in the prod datastore, however, is parent key + kind.  This is described here: http://code.google.com/appengine/docs/java/datastore/creatinggettinga... If it's important to you that the scope of the generated ids match between dev and and prod please file an issue.

                StringBuffer sb = new StringBuffer();                 KeyRange range = ds.allocateIds("a", 2);                 for (Key key : range) {                         sb.append("\n a " + key.toString());                 }                 range = ds.allocateIds("b", 2);                 for (Key key : range) {                         sb.append("\n b " + key.toString());                 }                 Key parentKey = KeyFactory.createKey("c", 1);                 sb.append("\n c " + parentKey.toString());                 range = ds.allocateIds(parentKey, "d", 2);                 for (Key key : range) {                         sb.append("\n d " + key.toString());                 }                 System.out.println(sb.toString());

The URL I posted earlier in the thread explains it, but here's a little bit more detail: A parent entity plus a kind defines an id-space, so entities with the same parent and the same kind are guaranteed to have unique ids.  For example, if you have an Entity with Parent:A Kind: Person Id: 10 you are guaranteed that no other entity with Parent A and Kind Person will be assigned an Id of 10.  However, an entity with a different Parent and Kind Person or an entity with Parent A and a different Kind _can_ be assigned an Id of 10.  The datastore pre-allocates batches of ids across multiple servers under-the-hood, so you can't make any assumptions about the Id that will get assigned in terms of contiguousness.  The only safe assumption is that the id will be unique for that Parent/Kind combination.

public void setProperty(java.lang.String propertyName, java.lang.Object value) Sets the property named, propertyName, to value. As the value is stored in the datastore, it is converted to the datastore's native type. This may include widening, such as converting a Short to a Long. All Collections are prone to losing their sort order and their original types as they are stored in the datastore. For example, a TreeSet may be returned as a List from getProperty(java.lang.String), with an arbitrary re-ordering of elements. Overrides any existing value for this property, whether indexed or unindexed. Note that Blob and Text property values are never indexed by the built-in single property indexes. To store other types without being indexed, use #setUnindexedProperty. Parameters:value - may be one of the supported datatypes, a heterogenous Collection of one of the supported datatypes, or an UnindexedValue wrapping one of the supported datatypes. Throws: java.lang.IllegalArgumentException - If the value is not of a type that the data store supports.