BeeSeek Blog

The alpha 2 version of Hive has been released and the main changes is the adoption of a modular structure. That means faster and easier development without adapting the main code every time, but it makes also upgrades and code reading easier and improves the stability. This also gives the possibility to personalize easily your own server writing your modules.

~ HoneyBee ~

Currently the only module developed already is HoneyBee, that is in alpha 1 version. This module collects all the statistics needed for the future page ranking.
But let's see how it saves the data in /hive/var/stats.log:

Extract sample from stas.log


Time: 1200939792
ID: 1196464515-54220160-65216570
MovTime: 4
StatTime: 18
Scrolls: 0
Url: http://127.0.0.1:50007/blog.beeseek.org/
Referer: http://localhost:50007/?query=beeseek%20blog


1. Time: records date and time of the search.
2. ID: identify the user who perform the search. The ID remains in the browser's cookies for 1 hour after the last use.
3. MovTime: records the time that the mouse moves.
4. StatTime: records the time that the mouse remains static.
5. Scrolls: records how many times the page was scrolled.
6. Url: records the address the stats are for.
7. Referer: records the address you came from.

~ Installation ~

To get the latest release you can use this page, where are available .deb packages and tarballs.

For Ubuntu users are also available our
repository:

deb http://ppa.launchpad.net/beeseek-core-devs/ubuntu/ hardy main devel
deb-src http://ppa.launchpad.net/beeseek-core-devs/ubuntu/ hardy main devel

Even if the packages are for Hardy, they are perfectly compatible with all the previous Ubuntu releases since Edgy. Anyway specific packages for Edgy, Feisty and Gutsy are going to be added too.

~ The future ~

The plans for the future are:
Hive: solve two blueprints (ipv6 e apport).
HoneyBee: set up the p2p network wich BeeSeek is going to be based on.

HoneyBee is the part of the project that is going to become our stable web application and will provide the BeeSeek service to everyone.

This is the HoneyBee's homepage:

As you can see, is the same as beeseek.org.
But have a look at the following screenshot: the results' page has just been updated to a new theme that reflect the style of our network:

All the restiling work was done in the BeeSeek Forum-it, where everyone can suggest new ideas to improve BeeSeek.

~ Installation ~

The page is now working installing Hive and HoneyBee as a module, using bazaar as already explained in the previous article Debugging.

Once downloaded both the programs, you need to copy all the items contained in the HoneyBee directory:

• ball-blue.png
• fireresult.html
• honeybee.py
• separator.png
• firepage.html
• head-bg.png
• honeybee.js
• logo.png

theese files (the list could change in the time because of the development activity) have to be copied to:

/hive/etc/modules

and the file:

• honeybee.conf

has to be copied in:

/hive/etc/

For all the Ubunutu users are also avalaible our repository:

deb http://ppa.launchpad.net/beeseek-core-devs/ubuntu/ DISTRO main devel

deb-src http://ppa.launchpad.net/beeseek-core-devs/ubuntu/ DISTRO main devel

In the place of DISTRO put the name of your distribution: Edgy, Feisty, Gustsy or Hardy.

~ Start up ~

Now is possible to use HoneyBee launching Hive (read here) and connecting your browser to:

http://localhost:50007

Happy BeeSeeking to everyone!

Today the prerelease of Hive Alpha 2 has been officially released, and added temporarily to BeeSeek releases.
Changelog: powerful and flexible settings manager, a great actions tracker, an almost perfect server platform. The package for debian-like platforms (soon to be together with its RPM version) offers an integration with pycentral and obeys, unlike alpha 1, to all new Debian policies.
This software, at the moment, only receives datas from the Firefox extension (soo to be started), only for statistical analysis. Datas are harvested and will be used to create our indexing algorithm. Prerelease will be tested soon, and we hope that the availability of
Alpha 2 will coincide with the end of tests. The code needs some clean up and a documentation update, we'll do our best to make the release available as soon as possible.

In case of slight delays, that additional time will be used to improve the software and correct eventual bugs.

Tarball (tar.gz): hive-0.2.01-prerelease.tar.gz
Tarball (tar.bz2): hive-0.2.01-prerelease.tar.bz2
Debian package (deb): hive-0.2.01-prerelease.deb

Translated by Ubuntista

Yesterday I interviewed the Team Leader of our Beeseek Developers
Team, andrea-bs, in the nice virtual parlour we have here :-)
He explained what he did to release Hive's first alpha. This is the
first of a long series of interviews we'll do.

JAK: Hi andrea-bs, what's your real name?

ANDREA-BS:Andrea Corbellini

JAK: The nice thing about internet is that everyone can work on projects like beeseek, comfortably from his home. So we'd like to know where you live...

ANDREA-BS: I live in Cremona, north Italy.

JAK: How did you learn about Beeseek?

ANDREA-BS: I regularly read Ubuntista's blog, he posted about this idea and I was immediately taken.

JAK: Why did you join a Beeseek?

ANDREA-BS: The main reason is that it is an open source project.

JAK: Why you decided to join a project that aims to enter a highly competitive market like search engines?

ANDREA-BS: Because I've always loved servers, p2p, databases and python, and this seemed a good chance to improve my skills and at the same time collaborate in a community to develop an open source software.

JAK: You are the Team Leader of the Developers Team, what do you know about that?

ANDREA-BS: I know Python pretty well, altough it never stops suprising me. I completed various projects that used homemade peer to peer networks. I never had a chance to delve into SQL, but I'm fine with that now.

JAK: Self learner?

ANDREA-BS: Yep, 100%!

JAK: Let's move on your work on Beeseek: at the moment you are working by yourself at the development of Hive. Can you elaborate on that, and tell us what you've accomplished so far?

ANDREA-BS: Hive is not a unique block, from a threading and code perspective, but is divided into different processes and moduls, each one regarding a specific task. The "Package settings" interprets and manages the other packages; because of its importance, it is the first one on which I focused my attention, and because it's the most complicate one. We need to be sure that the program is not at risk because of a bad configuration, for example. We included the logging component in the same package, to allow a single object to handle the most recurrent operations (harvesting settings and keeping track of actions). Most error handling was included in the logging part to reduce the number of calls to recurrent functions. The package that handles the server part, and so the socket, is clearly visible in the server package. This package gave us many problems, because the default libraries in Python, although very powerful, didn't allow the flexibility we needed. The "hive" file launches the other components, and checks if there is another instance already running, to avoid problems related to the sharing of resources. The executable Hive-Manager configures and manages Hive while it is running.

JAK: Today the alpha release of Hive has been published. What does this mean for Beeseek?

ANDREA-BS: Alpha 1 is an important step for Beeseek. If everything goes fine, we can collect datas and start diffusing the software.

JAK: so, there are new goals for the programming part?

ANDREA-BS: In the future I see a support for clients, that can really help the start of the project. The server must be capable of handling thousands of requests and answers without problems. I would also like to personally focus on Hive, rather than on Hive Manager. At this moment, hive-manager can be used, but is slightly difficult to address, and in few days we'll release a browser extension. I hope someone will help me soon in the future.

JAK: What problems you met to reach Alpha 1, and what the most fun parts?

ANDREA-BS: Nothing fun :D!
The main problem was with Dreamhost, that was killing the process whenever we tried to manage more than one connection. When I discovered this with a small stub, I changed many things in the
project, in fact the clients are not always connected, but ping every once in a while. The remaining was not difficult, and, if Alpha 1 will give us the expected results, that would be the nice moment I'm waiting for! There is one very interesting part, though, it is the overall project: I already worked on similar things, but I had to mix many skills and knowledges to obtain results in this case.

JAK: Well, we're at the end of this interview... If I forgot to ask you something, that's your moment now :D

ANDREA-BS: No, I don't think you forgot anything, today we have a new part to develop and I already want to write some code!

JAK: Well, we have two similar roles... we need to write things! Good work, and thank you!

Translated by Ubuntista

Released Hive and Hive-manager Alpha 1 Versions. Hive is the main application running on the BeeSeek server, while Hive-manager is a tool to control Hive from a local machine. Hive is now running with two instances on BeeSeek's server in order to test Hive itself. At the moment there are some problems, but we are analyzing them in order to understand if these problems are Hive bugs or hosting service problems.
Is active on the wiki a page dedicated to all the news about BeeSeek. The page is at http://www.beeseek.org/wiki/it/News
Have a great BeeSeek expirience!
CohibO.