task.txt

# PYTHON Dev. Interview Task:

# Scenario: You need to provide a server that mimics an ETL (extract-transfer-load) process.
# The server will receive arrays of file names through an open port (8081) and use these files for
# a “fake”:

# load operation (no files exist).
# • The server will handle a maximum of 4 load requests simultaneously with 40 sec timeout each.
# • Each request, encoded as a JSON, contains an array of filenames.
# • These file names should be loaded into a table by appending their names to it*.
# • The server sends a heartbeat as TIMESTAMP, into a database of your choice every min (such
# as
# PostgreSQL) .


# • There’s a 10% chance a load contains a corrupt file**.


# Load*
# Each request coming in through the open port as a consists array of strings. These strings are
# supposedly file names. But, we aren’t loading actual files as they don’t exist. Instead, we’ll
# simply take their names and append them to a table in database. The table structure and name
# is per your choice.


# Corrupt file**
# Each load request has a 10% chance of having one corrupted file. This corruption prevents the
# file from being loaded.
# Before each request starts, you’ll “roll the dice” for this to corruption to occur. Implement it in any
# way you like, A couple of ideas are:
# Filenames for load are objects containing an attribute “corrupt” or
# filenames are within a 2-item tuple in which the second item is a Boolean corrupt/not-corrupt.
# The goal is to catch an exception during the load operation and find which file caused the
# exception. Remember that, ultimately, we want to load all good files.
# Pay attention to coding standards. Add appropriate comments, keep the code readable and, if
# all else fails, import this.


# Good luck and have fun.