I’m not here.

I just released version 0.4 of pysmug and I’m pretty happy with the additional functionality.

The most significant additions to the ChangeLog referenced in my previous post are the migration of examples to tests and the overhauled configuration file format for pysmug.login().

tests

I’m generally a proponent of testing but since so much of pysmug is dependent on SmugMug services I added examples rather than tests. That’s changed with v0.4 — the examples have either been converted to code in SmugTool and/or migrated to a test case. I took the opportunity to learn nose and I’m not upset with the decision. When I switched from in-function preparing of SmugMug-style parameters names to using a decorator it was nice to have the tests to back me up.

configuration file format

The original configuration file format was evaled Python code. I was never perfectly happy with this approach and when I needed to add the ability to choose the class to instantiate in pysmug.login() I took the opportunity to change it. You can read more about the changes here.

When we lived in Chicago, my wife and I would spend hours in the Fern Room at the Garfield Park Conservatory during the depths of winter as a cheap way to get some heat and humidity. Now we live in the PNW, where we have ferns everywhere, and I don’t look at them the same way anymore. After two days of pruning and a number of piles of refuse, we have finally removed the old leaves on about, oh, a quarter of the ferns in the main part of the yard — a shitload [1] of work. We also cleared some hold outs from this winter’s English Ivy removal party with my dad. The yard looks much better.

[1] From Urban Dictionary:

The Shitload first became common usage during World War II. Farmers, who’s sheds were previously full of shit were called upon to empty their shit-filled barns in order to allow for the making of weapons.

A Shitload refers to between 1 and 4 tons of equine or bovine excrement (that’s horse and cowshit to you and me). More specifically Shitload describe the overfullness of the trailers used by farmers.

I’ve got a shitload of hay in my barn.

by Graham Nigel Goddard-Hill The West Country Jul 15, 2005

While I hammer away at erlsmug and try to figure out why it’s so much slower than pysmug and PycURL, I’m ready to release version 0.4 of pysmug.

Among the biggest changes are compatibility with new security changes required for version 1.2.2 of the SmugMug API. It’s now also possible to register function callbacks for PycURL’s verbose and progress handlers.

In testing erlsmug’s concurrency, I wrote a function to return all the details of an album as well as the details about each image, including EXIF. This requires 1 call to get the album info, 1 call to get the images for an album and one call each for simple and EXIF details per image. This function, albums_details was so useful I wanted it for pysmug.

Using batchables with pysmug, all image calls are handled in parallel and asynchronously for a pretty quick response. The implementation of the function is a great example of batchables in action. I can’t get the Erlang version to perform anywhere near as quickly.

In testing the new 1.2.2 API, it seems the images_upload method doesn’t require the *Key attribute though all other methods do appear to require it. I need to confirm this with SmugMug.

Changelog

The tag hasn’t been applied yet but I suspect this will be more or less the Changelog.

• fully tested with the 1.2.2 API changes
• added the ability to choose the API version to use (defaults to 1.2.2)
• added albums_details which returns detailed information about an album and all containing images (include EXIF if requested)
• added verbose and progress callbacks to assist in debugging
• do not send any keywords to SmugMug with a value of None
• updated to use hashlib for md5
• examples now either use anonymous logins or require a conf file
• added and fixed documentation

I’ve recently started a thread about editing EXIF GPS tags in my images from Expression Media. Microsoft has taken the stance EXIF data is read-only and GPS tags fall within this category. The referenced Wikipedia link lists a number of problems with editing EXIF data so maybe this isn’t entirely unwarranted.

The majority of tags edited post-capture are handled by IPTC but unfortunately the GPS set of tags reside solidly in EXIF. Since I can’t update my geo data from within EM I can, and do, edit it with other tools, namely PictureSync or the amazing ExifTool.

So who’s right? If EXIF data is meant to be static, and for the most part does contain static data, then MS has a somewhat sound argument. But if the rest of the industry eschews this notion then should MS really decide otherwise and inconvenience its user base?

Standards work best when they are.

My daughter loves ice cream too, having devoured an ice cream sandwich the very first time she ever saw one. Since one of the joys in my life is sharing ice cream with my family we have it quite frequently, sometimes at home and sometimes at Mora as a treat. If you find yourself at Mora, skip the vanilla (though it’s my favorite flavor theirs is nothing special) and try one of the more unusual varieties (my current choice being banana split).

Mora is quite expensive so 9 times out of 10 it’s either Alden’s or homemade at home, vanilla both. In an effort to cut back on quantities with my daughter we’ve taken to eating our ice cream from our espresso demitasse. This is great both because it’s still wonderfully satisfying to eat ice cream regardless of portion size and it’s a natural for espresso over the ice cream, which is without question the best ice cream topping known to man!