Сообщения

The project PSM (zero) is complete and needs you

Изображение
Yesterday, Pavel Stehule finished work on the implementation of the procedural language SQL/PSM to PostgreSQL. At the moment, the language supports all you need: the the simple things, arrays, composite types (composites), triggers; the additional function that returns a table, IN/OUT parameters; the features SQL/PSM — warning, exception handlers (in most on the basis of SQLCODE), a SIGNAL or RESIGNAL statement; some of the features of DB2 and MySQL — multi assign operators, support for magic SQLSTATE and SQLCODE variables. Some examples: the create or replace function test74_2() returns text as $$ begin atomic declare not_found condition for sqlstate '03000'; declare undo handler for not_found begin declare xx, yy text; get stacked diagnostics xx = condition_identifier, yy = returned_sqlstate; return xx || 'Signal is handled' || yy; end; signal not_found; end; $$ language psm0; create or replace function test66(a int, r out int) as $$ begi...

PageRank and the fight against cancer

Изображение
German scientists are using a modified version of PageRank to scan the tumors. This algorithm, NetRank, scans genes and proteins in cells and determines all the features of the relations between them, in the same way as PageRank scans the web page for hyperlinks. This approach can help in finding new ways of treating tumors. Christoph winter, doctor of medical Sciences, Lund University, Sweden: — First, we strongly experimented and tried to invent our own algorithm for finding the links. Then I decided that it is not necessary to reinvent the wheel when there are PageRank. While the study focused on tumors of the pancreas, the most common form of which is, pancreatic ductal adenocarcinoma, are causing more than 130,000 deaths per year (Europe and US). The researchers used NetRank to analyze 20000 proteins to determine the approximate parameters of their survival. Then it was identified 7 proteins that can help in the assessment of the "aggressiveness...

Create a tester for load testing of PostgreSQL

the Idea of this proektika (it "proektika") arose spontaneously. The company uses memory DB TimesTen contains a single large table with data more than 150 million records, and the volume is about 15 gigs. TimesTen always worked fine, the response for any request received within milliseconds, everyone is happy. One of these days, T10 was responding very long, the response time is increased to 3-5 seconds. Tehpotdergka of course started work on finding the problem, but in parallel we asked the question, why do T10 is used, why not move the base on a conventional RDBMS Oracle or Postgres ? needed to find out to carry out the appropriate tests. In the end, a little digging on the Internet frišnyh for testing was not. In the end of the day "on my knees" was written a small console utility which would have measured the response time of DBMS on different types of queries, collecting statistics, and in addition would also multithreaded, so load testing was the most o...

The creation of a search engine, or automation of Yandex.Server

Изображение
I'm from Belarus, the largest Internet service provider is byfly. This provider provides its users with free access to all sites that are hosted within the country (guest resource). Each user byfly in the Arsenal was a set of files with links to free resources available. So came the idea of creating a search engine for these resources, and in August 2009 it saw the first users. Attendance grew quite quickly and, at the peak of popularity, a resource was visited by about 34 000 unique users a day. In the heart of the search engine Yandex works.Server. Here is it the control panel: Once it became clear that having only this functionality (3 buttons is on/off search, on/off indexing and off Ya Server) sane search engine not to do. So I had a bit of work and it turned out: There have been numerous additional functions, such as: the the split index into resources and create for each of them, his mask and the rules of the index; the system automati...

9facts: debriefing

Изображение
/ > In mid-March, we actually shut down our startup 9facts.com , which I wrote on Habrahabr in December. And so by may I was ready to write this post. Start with the most important: the What mistakes do we? the 1. The idea was not validated using past The most important hypotheses for any startup: the the Product will use the following categories of users: ... the market size of product: ... Each of them needs to be tested. The most reliable method of this test — development prototype, everything else (surveys, etc.) has a much lower accuracy. We actually missed this step and decided to make a relatively quality product from. It happened because I (and not only) was the absolute confidence that we are doing the product that will find its users. In fact it was just a hypothesis. It is worth saying that the signals that the idea is risky — was: the the Not found any successful startup that uses a similar model describing the activities of ...

Improving search relevance in Sphinx

Изображение
Sphinx is a search engine for fast fulltextsearch, can obtain data from mysql, oracle and mssql can be itself a repository(realtime indexes). Sphinx also has a mode of operation through the api and via sphinxql — analog Protocol sql(with some restrictions), enabling you to connect to the sphinx search through the website with minimal code changes. This is one of the few great, large and open projects developed in Russia. In my life I saw how sphinx handles 100-200 search queries on 2 million records from mysql and the server is freely breathing and not throwing up, mysql starts to die already 10 requests per second on a similar config. The main problem of the documentation of sphinx in my opinion a small number of examples for most interesting settings, today I will try to tell in the examples about them. Options which I'll touch on this concern the algorithms and variations of the search. Everyone who works closely with sphinx not learn anything new, and newcomers will hop...

Report on participation in Seedcamp Berlin

Изображение
Recently on Habre we published a article about the experiences with the publication on TechCrunch and other articles on international experiences of Russian startups. I also want to share with abrasheva the experience that we received at Seedcamp'e , proshedshem this week in Berlin, moreover, that such a level of events for startups in Russia yet and positive emotions is through the roof. It all started with the fact that 3 weeks ago, I stumbled on information about the event, now even I do not remember where, and decided that we needed to try our strength. I had the idea of Seedcamp accelerator as the #1 in Europe, but especially over their actions never followed. Immediately after deciding what to try, I started filling it out, especially since before the time of enrollment was only a couple of days. The questionnaire consisted of a fairly standard for this kind of activities set of questions about the founders of the company, our strong traits and features. Some of ...