In 2012, Dan Pinto faced an unusual problem: he needed to buy a printing press. You can’t buy such
equipment on Ebay — to find a machine at the best price, you would have to go through hundreds of
specialized sites.
data and parse everything into a single database. This is how the idea of a startup selling complex
equipment was born — Machinio .
From idea to working product
Together with long-time friend Dmitry Rokhfeld, Dan launched the startup Machinio, which
systematizes information from several sources and helps find a suitable equipment seller from
thousands of offers.
The platform provides
a channel for communication between sellers and buyers to discuss details and clarify
specifications.
The first version of the product was written by Dan himself. As the service grew and the number of
Dan realized that it was difficult to combine two roles – managing the company and
maintaining the code in working order. During this period, the team turned to Evrone
to strengthen the in-house development team and expand the capabilities of the
service. We worked
on the server part, the Machinio core.
How does data collection work?
The Machinio database is formed automatically. Web crawlers go through sellers’ websites and
listings to collect the necessary data. The seller only signs a contract with the platform and pays for
the subscription.
source web driver created with the participation of Evrone. With its help, it is possible to collect data
from sites on React or Vue.
Ferrum and one crawler are enough
to parse the necessary information.
its own structure and design, the sites often have similar markup. That’s why we wrote just one
crawler and adapted it for different listings.
Data collection and processing takes place in two stages:
Then the goods are distributed by categories and spam number database models, for this, machine learning and a separate
team of developers are involved. There are positions that ML cannot handle . For
example, once an
entire plant and all the equipment available on it were displayed on the site. At such moments, the
participation of a specialist is necessary.
Parsing is done daily, but not in real time. prepare your site for voice search The market for special equipment is slow, and there is
The Future of Machinio
The result of their collaboration was a multifunctional service, the team achieved good results and
In belgium numbers 2018, Dan and
Dmitry decided to sell Machinio to a
large American network, Liquidity Services.
Ferrum is a new solution. It is an open source driver for managing Chrome via
WebSocket using the Chrome DevTools protocol. The driver’s performance exceeds
that of Selenium, Ferrum provides a high-level API to the browser and does not require additional software.
Features and usage scenarios
Ferrum is easy to use. By default, the driver goes in headless mode, but this can be changed in the settings. In addition to headless, there are several attributes to
customize the driver’s operation.
All you need to start testing is a driver, Chrome or Chromium. Ferrum connects to the
browser via the Chrome DevTools Protocol (CDP) and does not depend on Selenium, WebDriver or ChromeDriver. Chrome itself allows you to perform many operations
that are practically not supported by WebDriver.