Details about graded tasks

Write a script (shell pipe) that reads web server logs and prints the day with the highest number of requests.

The logs will use the same format as in the example from the labs:

2023-01-05,02:35:05,192.168.18.232,622,https://example.com/charlie.html
2023-01-05,09:01:33,10.87.6.151,100,https://example.com/bravo.html
2023-01-06,17:25:17,52.168.104.245,1033,https://example.com/delta/delta.html

For the above data we expect that the script will print the following on its output.

2023-01-05

The web log will come on standard input, print the date on standard output. You can safely assume that the input is not corrupted. Do not make any assumption about the order of the input lines.

You will be expected to perform the following tasks in Git from the command-line (some might be required to execute on the remote machine linux.ms.mff.cuni.cz).

• Configure your Git environment (author and e-mail)
• Clone a repository (from gitolite3@linux.ms.mff.cuni.cz or from GitLab or generic HTTPS)
• Create a commit
• Create a branch
• Switch between branches
• Merge a branch (and solve any conflicts)
• Push changes (branches) to server

Ensure that you can clone from gitolite3@linux.ms.mff.cuni.cz when using the school machines. Only authentication via public key will be available (i.e. upload your keys to 05/key.[0-9].pub files in your repository before the exam as explained in Lab 05).

Update: we have added a new CI job to your GitLab repositories that would warn you about typical issues with your keys. Feel free to execute locally via ./bin/run_tests.sh misc/keys.

The following command (replace LOGIN with your SIS/GitLab login in lowercase) will check that.

ssh -o ForwardAgent=no LOGIN@u-pl1.ms.mff.cuni.cz ssh -T gitolite3@linux.ms.mff.cuni.cz

You should see something like the following in the output:

hello LOGIN, this is gitolite3@linux running gitolite3 3.6.13-2.fc39 on git 2.44.0

 R W    lab05-LOGIN
 R      lab07-group-sum-ng

If you see the following, your keys are not setup correctly.

(LOGIN@u-pl1.ms.mff.cuni.cz) Password:
Permission denied, please try again.
Permission denied, please try again.
gitolite3@linux.ms.mff.cuni.cz: Permission denied (publickey,password).

Feel free to store the URL gitolite3@linux.ms.mff.cuni.cz somewhere on the local disk in your $HOME so that you do not have to copy it manually during the exam.

Host exam
    Hostname linux.ms.mff.cuni.cz
    User gitolite3

The focus of the exam is on working with Git. You will not be required to write any script on your own but we will be working with a repository containing the following script for printing simple bar charts in the console. You will be required to make some small modifications (such as fixing typos) but we will always guide you to the right place in the code.

import argparse
import sys

def parse_config():
    args = argparse.ArgumentParser(description='Console bar plot')
    args.add_argument('--columns', default=60, type=int, metavar='N')
    return args.parse_args()

def load_input(inp):
    values = []
    for line_raw in inp:
        line = line_raw.strip()
        if line.startswith('#') or not line:
            continue
        try:
            val = float(line)
        except ValueError:
            print(f"WARNING: ignoring invalid line '{line}'.", file=sys.stderr)
            continue
        values.append(val)
    return values

def print_barplot(values, scale, symbol):
    for val in values:
        print(symbol * round(val / scale))

def main():
    config = parse_config()
    values = load_input(sys.stdin)
    if not values:
        sys.exit(0)
    coef = max(values) / config.columns
    print_barplot(values, coef, '#')

if __name__ == '__main__':
    main()

You will be expected to perform the following operations with Makefile (some might be required to execute on the remote machine linux.ms.mff.cuni.cz).

• Build a project (or its component) using make
• Understand existing Makefile
• Make updates to the Makefile
  • Fix list of dependencies
  • Fix command invocation for a particular target
  • Merge multiple rules into a pattern rule
  • Add a new target to Makefile with correctly derived dependencies

Ensure that you can clone from gitolite3@linux.ms.mff.cuni.cz when using the school machines (as for the Git CLI exam). See hints and details above.

There can be arbitrary number of teams in the tournament and each team can submit their implementation of the competitive task. The implementation is then evaluated through a set of automated tests and a short log is copied to a well known location.

Your task is to process the output of the automated tests and generate a summary web page. Because the actual generation of HTML is not interesting the task will stop at the boundary of generating a set of Markdown pages.

The input data will be stored in the tasks directory. Each subdirectory corresponds to one task, actual results are in files named by the team name with the .log.gz extension (i.e. it is a plain text file compressed with gzip). Each line in the log file either starts with pass (including the space) or fail or can be safely ignored for our purposes.

It is expected that the shell script can be (eventually) installed in some $PATH-like directory but it will always read data from the current working directory.

You are supposed to generate an overall ordering of the teams where each passed test (i.e., the line starting with pass ) counts as one point (and the points are summed across all tasks). And for each team you also need to prepare a page where a breakdown of points across each task is shown. This page will also link to the original log (you are expected to copy and decompress the log to the same directory where the markdowns are).

As an example, the input directory tree can look like this:

tasks/
├── m01
│   ├── alpha.log.gz
│   └── bravo.log.gz
└── m02
    ├── alpha.log.gz
    ├── bravo.log.gz
    └── charlie.log.gz

And the contents of tasks/m01/alpha.log.gz can be the following (recall that we are interested in pass/fail lines only) after decompression:

pass Test one
fail Test two
  Extra information about failure.
fail Test three

A complete dataset (together with expected output) can be downloaded from the examples repository.

Then we expect to generate the following index.md:

# My tournament

 1. bravo (5 points)
 2. alpha (3 points)
 3. charlie (1 points)

And for each team a special page like this:

# Team alpha

+--------------------+--------+--------+--------------------------------------+
| Task               | Passed | Failed | Links                                |
+--------------------+--------+--------+--------------------------------------+
| m01                |      1 |      2 | [Complete log](m01.log).             |
| m02                |      2 |      0 | [Complete log](m02.log).             |
+--------------------+--------+--------+--------------------------------------+

The output directory would then contain the following files:

out/
├── index.md
├── team-alpha
│   ├── index.md
│   ├── m01.log
│   └── m02.log
├── team-bravo
│   ├── index.md
│   ├── m01.log
│   └── m02.log
└── team-charlie
    ├── index.md
    ├── m01.log
    └── m02.log

Your script must accept -o argument for passing the name of the output directory and -t for specifying alternative index page title (instead of the default My tournament. Your tool must accept both variants of -odirname and -o dirname invocation for both options.

You can safely assume that team names and task names will be only valid C identifiers (i.e. only English alphabet and numbers without any special characters). You can safely assume that task names etc. will fit into the table above (regarding column widths).

You are expected to use temporary files to store intermediate results and you are expected to solve the task completely in shell.

You cannot use AWK, PERL, Python or any other language; using sed is allowed (but probably not needed). You cannot use any advanced Bash features such as (associative) arrays or [[ and (( constructs (use of $(( )), $( ) and test(1) is allowed, though). In other words, we expect you will solve the implementation using constructs and programs you have seen during the labs (and the task is such that you should be able to write it yourself without external help).

We expect you will use external tools to drive your implementation but you must understand the whole script before submitting it and you must mark all parts that were not authored by you personally (and if you are using tools such as ChatGPT, you must submit the whole log of your communication with the tool).

All Shellcheck issues must be resolved for the submission to be accepted.

Extension

Check if a config.rc exists and load it: it might override the title of the whole tournament and output to use different output directory then out.

For example, if config.rc contains the following lines then the title in top-level index.md would read # NSWI177 Tournament and generated files will be stored in output_md/ directory.

title="NSWI177 Tournament"
output="output_md/"

These values might still be overriden by command line switches (i.e. the ordering is default value hard-coded in the script, values in config.rc and -o and -t switches).

In each task/* subdirectory check for meta.rc file that may contain the line name=".." to use a different task name than the directory name. This name will be used instead in the table mentioned above. The name will not use any special characters except space and a dash.

The .rc files are expected to be sourcable to shell script and will contain valid shell constructs. You have to check that the file actually exists, though.

Submission and grading

Submit your solution into exam/t04 subdirectory in your NSWI177 repository.

Store the script into exam/t04/tournament.sh (do not split it into multiple files). When you copy fragments from sites such as StackOverflow we expect you will comment them directly in the tournament.sh. Communication with AI-driven sites store into ai.log file (plain text file with clearly marked portions with your input and with the answers).

If you create further testing datasets, feel free to store them into demo subdirectory (similarly to our inputs, e.g. demo/one/tasks and demo/one/output with expected output).

We provide only basic automated tests: feel free to extend them yourself. Note that diff can be used with -r to compare whole directory subtrees which might come useful when comparing the generated output with the expected one.

The grading of the task is only passed/failed (and passed with extension). When evaluating your implementation we will check that it passes basic functionality checks and then check the quality of your script by a combination of automated checks and manual inspection. This will include checks that the script removes temporary files after itself, does not remove existing files randomly etc. For the quality we will look at variable names, comments of the overall structure or decomposition into functions. The implementation will pass if the amount of such issues is within reasonable limits.

We intentionally do not provide a comprehensive checklist as that would lead to optimization against the checklist instead of thinking how to write a reasonable implementation.

The submission deadline is 2024-05-05.

We expect you will use external tools to drive your implementation but you must understand all the parts before submitting it and you must mark all parts that were not authored by you personally (and if you are using tools such as ChatGPT, you must submit the whole log of your communication with the tool).

Context

In this task you will be working with a simple Python project that is able to render Jinja templates.

As a trivial example (which you can also find in the examples subdirectory of the project repository) it will be able to perform the following transformation.

We will have the following array (list) in JSON:

[
  {
     "name": "Introduction to Linux",
     "code": "NSWI177",
     "homepage": "https://d3s.mff.cuni.cz/teaching/nswi177/"
  },
  {
     "name": "Operating Systems",
     "code": "NSWI200",
     "homepage": "https://d3s.mff.cuni.cz/teaching/nswi200/"
  }
]

We will have the following input file:

Our courses
===========

Below is a list of (almost) all of our courses.

And we will have the following template. Note that control structures of the template use {% (or {%- to strip surrounding whitespace) and {{ for variable expansion.

{{ content }}

{%- for course in data -%}
 * [{{ course.name }} ({{ course.code }})]({{ course.homepage }}) {{ NL }}
{%- endfor %}

When executed with our renderer, we will get the following output.

Our courses
===========

Below is a list of (almost) all of our courses.

* [Introduction to Linux (NSWI177)](https://d3s.mff.cuni.cz/teaching/nswi177/)
* [Operating Systems (NSWI200)](https://d3s.mff.cuni.cz/teaching/nswi200/)

Source code

The source code for the above implementation was already prepared for you and soon you will have access to a new project under teaching/nswi177/2024 subtree in GitLab where you will work.

The repository also contains several tests. There are unit tests in Python (using pytest) as well as higher-level tests (let us call them integration tests even that might be a bit overstated) written in BATS.

The invocation is described in the project README.

Base task

Your main task is to setup basic CI on GitLab for this project and prepare the project for distribution.

The CI must execute the Python unit tests and fail the job on any error. The preparation for distribution means that after your changes we will be able to install the templater via pip and have the templater available as nswi177-jinja-templater command.

In other words, the following commands would install the templater into a fresh virtual environment and the invocation of the last command would print a short help from our program (we assume we are in an empty directory that is not related in any way to any clone of your project).

python3 -m venv templater-venv
. ./templater-venv/bin/activate
pip install git+ssh://git@gitlab.mff.cuni.cz/teaching/nswi177/2024/t05-LOGIN
nswi177-jinja-templater --help

And for the CI we expect that your project would have a job unittests running the Python tests.

The CI should use the python:3.12-alpine image and should install the versions from requirements.txt file. We expect that pytest (and perhaps other test-related libraries) will not be mentioned in the requirements.txt but rather inside requirements-dev.txt file.

You will notice that the tests are not passing as some of the Jinja filters are not implemented.

Do not worry, though. Your coworker Alice already implemented them in her branch in her repository gitolite3@linux.ms.mff.cuni.cz:t05-alice.

Merge her implementation into your repository to get the full implementation.

If you do everything correctly, your CI job log would look like this (we have blurred the used commands for obvious reasons).

Optional extension

Optionally, you will extend the CI to also execute the BATS tests.

We expect that you will add a new job called integrationtests that installs the package (so that the nswi177-jinja-templater command is available) and run all the BATS files in the tests subdirectory (recall that simple pip install . should work like a charm).

You will need to install bats via apk add first (it is okay to install this package on every CI run).

Note that the current implementation in common.bash invokes the program via call to env PYTHONPATH=src python3 -m nswi177.templater – change this to call nswi177-jinja-templater instead. You will also need to replace --kill-after switch for the timeout command to plain -k as the long variant is not supported in Alpine.

The amount of BATS tests is quite low so you should also merge the work of your co-workers Bob and Charlie into your repository. Again, perform a normal merge and not a rebase when integrating their work.

Bob has his repository at gitolite3@linux.ms.mff.cuni.cz:t05-bob while Charlie put his copy to our website to https://d3s.mff.cuni.cz/f/teaching/nswi177/202324/t05-charlie.git.

In case of conflicts make sure you resolve them correctly – you certainly do not want to drop any of the tests.

Submission and grading

Submit your solution into the t05-LOGIN repository on GitLab.

We will check that the project can be installed via pip install git+https://gitlab.../t05-LOGIN/ and that your CI configuration is correct (after all the merges your jobs should be in the green). We will check that the Python code can be executed after pip install -r requirements.txt via python3 -m src.nswi177.templater (or via python -m nswi177.templater if the package itself is also installed) and that the Pytests can be run after installing from requirements-dev.txt.

When you copy fragments from sites such as StackOverflow we expect you will comment them directly in the appropriate sources, communication with AI-driven sites should be stored into exam/t05/ai.log file in your normal submission repository (student-LOGIN), again a plain text file with clearly marked portions with your input and with the answers.

The submission deadline is 2024-06-16.