Using the terminal

Bash¶

Default on most Linux systems.

Scripts – standard .sh.
Features – readline, globbing, job control, arrays.
Extensible – functions, aliases.

greet() { echo "Hello, $1!"; }
greet world

Zsh¶

Default on macOS.

Completion – advanced globbing (**/*.py).
Correction – auto-fixes typos.
Prompts – themes, colors.
Plugins – oh-my-zsh, prezto.

PROMPT='%F{green}%n@%m%f:%F{blue}%~%f $(git_prompt_info)%f '

Fish¶

Friendly Interactive Shell.

Syntax – uses set, no export.
Autosuggest – live history hints.
Errors – human-readable.
Config UI – fish_config.

set greeting "Hello, fish!"
echo $greeting

PowerShell¶

Cross-platform, object-based.

Objects – pipelines pass data.
Cmdlets – Get-Process, Get-Service.
Language – loops, try/catch.
Integration – all OSes.

Get-Process | Sort-Object CPU -Descending | Select-Object -First 5

Summary¶

Shell	Use	Strength
Bash	Scripts	POSIX, standard
Zsh	Interactive	Completion, themes
Fish	Interactive	Autosuggest, simplicity
PowerShell	DevOps	Object pipelines

Pick: Bash for scripts, Zsh/Fish for interactivity, PowerShell for cross-platform.

Bash¶

pwd        # show path
ls -la     # list files
mkdir dir  # make dir
cd dir     # change dir
cd ..      # up one
cd -       # previous dir

Files¶

touch f.txt      # new file
rm f.txt         # delete
rm -r dir        # remove dir
cp a b           # copy
mv a b           # move/rename

Permissions¶

chmod 644 file.txt

rw- r-- r--: owner read/write, others read. Use man chmod for more modes.

Searching and reading¶

find . -name "*.txt"
grep -nr "hello" .
cat hello.txt
less hello.txt

Scroll: Shift + Arrow, search: Ctrl+R.

Redirection¶

ls -l > out.txt   # write
echo hi >> out.txt  # append
1> stdout  2> stderr  &> both

history | grep "cd"

/dev/null discards output.

Processes¶

ps -A     # all
top -o %MEM
kill -9 PID

.bashrc / .zshrc¶

Run at login. Add aliases or settings.

alias pg='ping google.com'

Lab 01¶

Automating slide page generation¶

Assume you have cloned the course repository

git clone git@github.com:luca-heltai/sspa.git

Let’s explore the structure of the repository:

https://github.com/luca-heltai/sspa

Overview of the course page¶

The course materials are in jupyterbook/lectures/ and jupyterbook/slides/
The software jupyter-book (https://jupyterbook.org/) builds the course webpage from these (text) sources
It generates the static HTML page that is then hosted on GitHub Pages (https://luca-heltai.github.io/sspa/)

Problem statement¶

jupyterbook does not natively support slides
jupyterbook renders slides content as a single long html page (take a look at intro)
We would like to provide also an embedded (and interactive slideshow) in each lecture page

Solution¶

reveal.js (https://revealjs.com/) can render markdown slides as interactive slideshows

Problem:

We need to create a landing page for each slideshow (e.g., slides01.html) that reveal.js can use to render the slides from the markdown content (e.g., slides01.md)

Technical issue¶

jupyterbook generates its pages in a temporary folder during the build process
we need to copy the generated slidesXX.html there after jupyterbook has run
we want this to be automatic

Look at the Makefile, and see how the slides are copied after the book is built:

make build
make copy

Goal of today’s lab¶

Updates slide landing pages automatically.
Create slides (slidesXX.html) from content (slidesXX.md).
Embed slides in lecture notes (lectureXX.md).
Avoid manual errors
Allow easy updates to templates, styles, etc.

What¶

Inputs

Template HTML (from a reveal.js example): codes/lab_01/slides_template.html
Footer template (to add at the end of each lecture page): codes/lab_01/lecture_footer.md
Slide sources (the only content I want to write): jupyterbook/slides/slidesXX.md

Outputs

Landing pages: jupyterbooks/slides/slidesXX.html
Footer appended to: jupyterbooks/lectures/lectureXX.md

Mapping:

slides00.md → slides00.html → referenced in lecture00.md
slides01.md → slides01.html → referenced in lecture01.md
...

How¶

Find all jupyterbook/slides/slides*.md.
Extract XX.
Create slides/slidesXX.html by replacing slidesXX.md in the HTML template with the real filename (sed).
Build footer by replacing XX in lecture_footer.md.
If lectureXX.md does not reference /slides/slidesXX.html, append the footer.

Script: `codes/lab_01/generate_slides.sh`¶

generate_slides.sh

#!/usr/bin/env bash
#
# Generate HTML landing pages from Markdown slides,
# then ensure each lecture file ends with the correct footer block.
#
# Usage:
#   ./generate_slides.sh [--slides-dir PATH] [--lectures-dir PATH] [--templates-dir PATH] [--output-dir PATH] [-n] [-v]
#
# Defaults match the instructor’s layout.

set -euo pipefail

# Defaults (adjustable via flags)
SLIDES_DIR="jupyterbook/slides"
LECTURES_DIR="jupyterbook/lectures"
TEMPLATES_DIR="codes/lab01"
OUTPUT_DIR="jupyterbook/slides"
BASE_URL=""
DRY_RUN=0
VERBOSE=0

log() { printf '%s\n' "$*" >&2; }
vlog() { [ "$VERBOSE" -eq 1 ] && log "$@"; }

# Parse flags
while [ $# -gt 0 ]; do
  case "$1" in
    --slides-dir)     SLIDES_DIR="$2"; shift 2;;
    --lectures-dir)   LECTURES_DIR="$2"; shift 2;;
    --templates-dir)  TEMPLATES_DIR="$2"; shift 2;;
    --output-dir)     OUTPUT_DIR="$2"; shift 2;;
    --base-url)       BASE_URL="$2"; shift 2;;
    -n|--dry-run)     DRY_RUN=1; shift;;
    -v|--verbose)     VERBOSE=1; shift;;
    -h|--help)
      sed -n '1,30p' "$0" | sed 's/^# \{0,1\}//'
      exit 0;;
    *) log "Unknown option: $1"; exit 2;;
  esac
done

HTML_TEMPLATE="$TEMPLATES_DIR/slides_template.html"
FOOTER_TEMPLATE="$TEMPLATES_DIR/lecture_footer.md"

# Checks
[ -f "$HTML_TEMPLATE" ]   || { log "Missing template: $HTML_TEMPLATE"; exit 1; }
[ -f "$FOOTER_TEMPLATE" ] || { log "Missing template: $FOOTER_TEMPLATE"; exit 1; }
[ -d "$SLIDES_DIR" ]      || { log "Missing slides dir: $SLIDES_DIR"; exit 1; }
mkdir -p "$OUTPUT_DIR"

# Create HTML from template for each slidesXX.md
shopt -s nullglob
slides=( "$SLIDES_DIR"/slides*.md )
if [ ${#slides[@]} -eq 0 ]; then
  log "No slide sources found in $SLIDES_DIR (expected files like slides00.md)."
  exit 0
fi

for md in "${slides[@]}"; do
  md_base="$(basename "$md")"                 # slidesXX.md
  num="${md_base#slides}"                     # XX.md
  num="${num%.md}"                            # XX
  html_out="$OUTPUT_DIR/slides${num}.html"    # slides/slidesXX.html
  lecture_md="$LECTURES_DIR/lecture${num}.md" # jupyterbooks/lectures/lectureXX.md

  vlog "Processing $md_base -> $html_out; lecture file: $(basename "$lecture_md")"

  # 1) Generate HTML landing page from template, replacing placeholder
  # Replace both 'slidesXX.md' and 'BASEURL' placeholders.
  if [ "$DRY_RUN" -eq 0 ]; then
    sed -e "s/slidesXX\.md/$md_base/g" \
        -e "s|BASEURL|$BASE_URL|g" \
        "$HTML_TEMPLATE" > "$html_out"
  fi
  vlog "Wrote $html_out"

  # 2) Ensure lecture footer exists and is up to date
  if [ ! -f "$lecture_md" ]; then
    log "Lecture file not found, skipping footer: $lecture_md"
    continue
  fi

  # Build the concrete footer block by substituting XX with the numeric token.
  # This covers both iframe and link occurrences.
  # Use 'printf %s' to preserve newlines exactly.
  footer_block="$(sed -e "s/XX/$num/g" -e "s|BASEURL|$BASE_URL|g" "$FOOTER_TEMPLATE")"

  # Check if the lecture already references the correct HTML slide URL.
  # If it does, we assume the footer is present.
  if grep -Fq "slides${num}.html" "$lecture_md"; then
    vlog "Footer already present in $(basename "$lecture_md")"
    # Remove any existing footer block to replace it with the updated one.
    sed -e '/<!-- FOOTER START -->/,/<!-- FOOTER END -->/d' "$lecture_md" > "${lecture_md}.tmp"
    mv "${lecture_md}.tmp" "$lecture_md"
    vlog "Removed old footer in $(basename "$lecture_md")"
  fi

  vlog "Appending footer to $(basename "$lecture_md")"
  if [ "$DRY_RUN" -eq 0 ]; then
    # Ensure the file ends with a newline, then append a blank line and the footer.
    # Also insert a horizontal separator if the file already has content.
    if [ -s "$lecture_md" ]; then
      sed '${/^$/d;}' "$lecture_md" > "${lecture_md}.tmp"   # remove last blank line if present
      mv "${lecture_md}.tmp" "$lecture_md"
      printf '\n' >> "$lecture_md"
    fi
    printf '%s\n' "$footer_block" >> "$lecture_md"
  fi
done

log "Done."

Run the script¶

Make sure it is executable:

chmod +x codes/lab_01/generate_slides.sh

Run it:

./codes/lab_01/generate_slides.sh -v
# or dry-run first:
./codes/lab_01/generate_slides.sh -n -v

Breaking it down¶

Shell navigation and file ops
Globbing and pattern matching
sed for templating
grep for checks
Safe scripting patterns

Project layout¶

Expected folders¶

codes/lab_01/
  ├─ slides_template.html
  ├─ lecture_footer.md
  └─ generate_slides.sh   # we will look at this
jupyterbook/slides/
  ├─ slides00.md
  ├─ slides01.md
  ├─ ...
  ├─ (the script will write slidesXX.html here)
jupyterbooks/lectures/
  ├─ lecture00.md
  ├─ lecture01.md
  └─ ...

Creating the templates¶

mkdir -p codes/lab_01

mkdir -p creates parents as needed.
Verify:

ls -la codes/lab_01

Templates¶

HTML template snippet¶

The file codes/lab_01/slides_template.html must include (this is how reveal.js loads markdown):

...
<section 
  data-markdown="slidesXX.md" 
  data-separator="\n----\n" 
  data-separator-vertical="\n---\n">
</section>
...

XX is a placeholder we will replace.
\n is a newline (in regex).

codes/lab_01/lecture_footer.md:

<iframe src="/slides/slidesXX.html" width="100%" height="800px" style="border: none;"></iframe>

```{admonition} 🎬 View Slides
:class: tip

**[Open slides in full screen](/slides/slidesXX.html)** for the best viewing experience.
```

Again, XX is a placeholder.

Scripting strategy¶

Inputs → Outputs¶

Input: each jupyterbook/slides/slidesXX.md
Output A: jupyterbook/slides/slidesXX.html from HTML template by replacing slidesXX.md
Output B: ensure jupyterbook/lectures/lectureXX.md ends with a footer referencing /slides/slidesXX.html

Why a script¶

Repeatable for any number of slides
Avoid manual errors
Keep lectures and slides in sync
Easy to update templates/styles later

Build the script¶

Create the file¶

cd codes/lab_01
touch generate_slides.sh
chmod +x generate_slides.sh

touch creates an empty file.
chmod +x makes it executable.

Define defaults¶

These are variables:

SLIDES_DIR="jupyterbook/slides"
LECTURES_DIR="jupyterbooks/lectures"
TEMPLATES_DIR="codes/lab_01"
OUTPUT_DIR="jupyterbook/slides"

Keep paths configurable.
We’ll add flags to override.

Variables¶

Assign with =, no spaces.
Reference with $VAR or ${VAR}.
Quote variables: "$VAR" to avoid word-splitting.

Operations on variables¶

basename "/path/to/slides01.md"    # slides01.md
num="${md_base#slides}"            # remove prefix → 01.md
num="${num%.md}"                   # remove suffix → 01
echo "Number is $num"              # prints "Number is 01"

The ${VAR#prefix} and ${VAR%suffix} syntax removes parts.
Useful for extracting tokens from filenames.
echo prints to stdout and expands variables.

Loop over slide sources¶

shopt -s nullglob
for md in "$SLIDES_DIR"/slides*.md; do
  md_base="$(basename "$md")"     # slidesXX.md
  num="${md_base#slides}"         # XX.md
  num="${num%.md}"                # XX
  echo "Found $md_base with number $num"
done

Globbing picks only matching files.
Parameter expansion extracts the number.

Loops¶

shopt -s nullglob         # avoid errors if no matches
for item in list; do
  # commands using $item
done

Iterates over each item in the list.
Use shopt -s nullglob to avoid errors if no matches.

What is a “list”?

1. Explicit values¶

for n in 1 2 3 4 5; do
  echo "Number: $n"
done

2. Brace expansion¶

for n in {01..05}; do
  echo "Slide $n"
done

3. “Globbing” over files¶

for f in slides/*.md; do
  echo "Found file: $f"
done

4. Command substitution¶

for user in $(cat users.txt); do
  echo "User: $user"
done

5. Array elements¶

arr=(red green blue)
for color in "${arr[@]}"; do
  echo "Color: $color"
done

6. Slices within arrays¶

arr=(one two three four five)
for item in "${arr[@]:1:3}"; do
  echo "Item: $item"
done

Generate `slidesXX.html` with `sed`¶

sed -e "s/slidesXX\.md/$md_base/g" \
    -e "s/slideXX\.md/$md_base/g" \
    "$TEMPLATES_DIR/slides_template.html" \
    > "$OUTPUT_DIR/slides${num}.html"

-e can stack multiple substitutions.
We guard both slidesXX.md and a possible slideXX.md typo.
Always quote variables.

What is `sed`¶

sed = stream editor
Processes text line by line, applying editing commands.
Useful for find-and-replace, insertion, deletion, or pattern-based edits.

Reads from:
- standard input (echo "text" | sed ...)
- or files (sed ... file.txt)
By default, prints the modified text to standard output.

Basic syntax¶

sed 's/pattern/replacement/' file.txt

s = substitute command
Replaces the first match of pattern per line.
Add g to replace all matches in each line:

sed 's/foo/bar/g' file.txt

Output goes to the terminal. Use redirection to save it:

sed 's/foo/bar/g' file.txt > new.txt

Common examples¶

# Replace all "cat" with "dog" in file
sed 's/cat/dog/g' animals.txt

# Replace only on specific lines (line 1 to 5)
sed '1,5 s/error/warning/g' log.txt

# Delete blank lines
sed '/^$/d' notes.txt

# Show only matching lines (like grep)
sed -n '/TODO/p' script.sh

/pattern/ limits action to lines that match.

Using variables and escaping¶

When using variables, wrap the expression in double quotes:

name="Alice"
sed "s/USER/$name/g" template.txt > output.txt

To match special characters, escape them:

sed 's/file\.txt/archive.txt/' list.txt

The option -i edits files in place (careful!):

sed -i 's/old/new/g' file.txt

In macOS (BSD sed), avoid -i unless you supply a backup suffix:

sed -i '' 's/foo/bar/g' file.txt   # macOS
sed -i 's/foo/bar/g' file.txt      # Linux

Use sed in scripts for reliable, repeatable text transformations.

lecture_md="$LECTURES_DIR/lecture${num}.md"
footer_block="$(sed "s/XX/$num/g" "$TEMPLATES_DIR/lecture_footer.md")"

if grep -Fq "/$OUTPUT_DIR/slides${num}.html" "$lecture_md"; then
  echo "Footer already present in $(basename "$lecture_md")"
else
  printf '\n%s\n' "$footer_block" >> "$lecture_md"
  echo "Appended footer to $(basename "$lecture_md")"
fi

grep -Fq does a quiet fixed-string check.
If not found, append the concrete footer.

Run and verify¶

First a dry run¶

./codes/lab_01/generate_slides.sh -n -v

Confirms which files would be written.

Execute¶

./codes/lab_01/generate_slides.sh -v

Check outputs:

ls jupyterbook/slides/ | grep '^slides[0-9][0-9]\.html$'

Spot-check an HTML file¶

head -n 40 slides/slides00.html

Confirm it references slides00.md in the data-markdown attribute.

tail -n +1 jupyterbooks/lectures/lecture00.md | sed -n '$-40,$p'

Ensure the iframe and the “Open slides in full screen” link show /slides/slides00.html.

Key commands explained¶

pwd      # print working directory
ls -la   # list with details, including hidden files
cd path  # change directory
cd -     # jump back to previous directory

Globbing and patterns¶

ls jupyterbook/slides/slides*.md
# * matches any string
# ? matches a single char
# [0-9] matches a digit

Use quotes around variables to avoid word-splitting.

`sed` substitutions¶

sed 's/OLD/NEW/g' input > output
# Escaping literal dots:
sed 's/slidesXX\.md/slides07.md/g' template.html > out.html

g means “replace all occurrences in a line.”

`grep` checks¶

grep -Fq "/slides/slides07.html" jupyterbooks/lectures/lecture07.md
# -F fixed string, -q quiet (exit code only)
echo $?  # 0 → found, 1 → not found

Appending text safely¶

printf '\n%s\n' "$footer_block" >> jupyterbooks/lectures/lecture07.md

printf preserves newlines and avoids echo portability issues.

Troubleshooting¶

Common issues¶

Wrong paths: use -v and echo variables.
No matches: verify filenames are slidesXX.md with zero-padded numbers.
macOS sed: we used portable syntax, no -i in place.
Permissions: chmod +x generate_slides.sh.

Re-run anytime¶

The script is idempotent for footers. It only appends when the link to the exact slidesXX.html is missing.

Recap¶

What we automated¶

HTML landing pages from a template
Lecture footers that embed and link to the slides
Safe checks to avoid duplicates

Takeaway¶

Use Bash to codify repeatable edits:

glob files → extract tokens
sed for templating
grep to check state
append only when needed

Ready to integrate into your build process or CI.

Bash¶

Zsh¶

Fish¶

PowerShell¶

Summary¶

Bash¶

Basic navigation¶

Files¶

Permissions¶

Searching and reading¶

Redirection¶

Processes¶

.bashrc / .zshrc¶

Lab 01¶

Automating slide page generation¶

Overview of the course page¶

Problem statement¶

Solution¶

Technical issue¶

Goal of today’s lab¶

What¶

How¶

Script: codes/lab_01/generate_slides.sh¶

Run the script¶

Breaking it down¶

Project layout¶

Expected folders¶

Creating the templates¶

Templates¶

HTML template snippet¶

Footer template¶

Scripting strategy¶

Inputs → Outputs¶

Why a script¶

Build the script¶

Create the file¶

Define defaults¶

Variables¶

Operations on variables¶

Loop over slide sources¶

Loops¶

1. Explicit values¶

2. Brace expansion¶

3. “Globbing” over files¶

4. Command substitution¶

5. Array elements¶

6. Slices within arrays¶

Generate slidesXX.html with sed¶

What is sed¶

Basic syntax¶

Common examples¶

Using variables and escaping¶

Check and append footer¶

Run and verify¶

First a dry run¶

Execute¶

Spot-check an HTML file¶

Spot-check a lecture footer¶

Key commands explained¶

Navigation¶

Globbing and patterns¶

sed substitutions¶

grep checks¶

Appending text safely¶

Troubleshooting¶

Common issues¶

Re-run anytime¶

Recap¶

What we automated¶

Takeaway¶

Script: `codes/lab_01/generate_slides.sh`¶

Generate `slidesXX.html` with `sed`¶

What is `sed`¶

`sed` substitutions¶

`grep` checks¶