DEV Community: Steven Keil

Finding Email Attachments Yourself with IMAP Search (and When You'd Rather Not)

Steven Keil — Sun, 25 Oct 2026 13:16:52 +0000

If you only need last month's receipts, you don't need a tool. Gmail's search operators and a few lines of Python will get you there.

Start with Gmail search syntax

In the Gmail search bar:

has:attachment filename:pdf after:2026/09/01 before:2026/10/01

Combine from:, subject: and has:attachment to narrow things down, e.g. from:billing@example.com has:attachment.

Pull attachments via IMAP

If you want them saved locally without clicking through each email, Gmail supports IMAP with an app password (this requires 2-step verification to be enabled first). Note that some company (Workspace) accounts block app passwords through an admin policy. A minimal script:

import imaplib, email, os

mail = imaplib.IMAP4_SSL("imap.gmail.com")
mail.login("you@gmail.com", "your-app-password")
mail.select("inbox")

# Gmail search extension via IMAP
status, data = mail.search(None, 'X-GM-RAW "has:attachment after:2026/09/01"')

for num in data[0].split():
    status, msg_data = mail.fetch(num, "(RFC822)")
    msg = email.message_from_bytes(msg_data[0][1])
    for part in msg.walk():
        if part.get_content_disposition() == "attachment":
            filename = part.get_filename()
            if filename:
                safe_name = os.path.basename(filename)  # never trust a path from an email
                with open(safe_name, "wb") as f:
                    f.write(part.get_payload(decode=True))
                print("Saved:", safe_name)

mail.logout()

This works for Gmail and any other IMAP mailbox (Outlook, Yahoo, iCloud, GMX, web.de) if you swap the host and use that provider's IMAP settings.

A few gotchas before you rely on this for anything recurring

Re-running the script will save the same file twice unless you add your own dedupe logic.
Tiny inline images (signature logos) show up as "attachments" too and will clutter your folder.
Two attachments with the same name overwrite each other.
If you want things sorted into dated folders automatically, or want to keep it running in the background so new invoices get saved the moment they arrive, you'll need to build that layer yourself.

If you'd rather skip writing and maintaining that part

I made a small tool that does exactly this. It filters by sender, subject, Gmail search, date and file type, skips signature images and tiny files, never saves the same file twice, sorts everything into dated folders like 2026/10/2026-10-01_Invoice 42.pdf, and has a watch mode so it keeps running. It works with Gmail (app password and 2-step verification) and any IMAP mailbox, and is read-only by default.

It is mine and it is paid, 12 EUR: https://skeilster.gumroad.com/l/gmail-attachment-saver

Finding Exact Duplicate Photos Without Installing Anything

Steven Keil — Sun, 25 Oct 2026 12:55:53 +0000

If your photo library has grown out of control, there's a good chance a chunk of it is exact duplicates: the same file copied during a backup, a sync, or an import you ran twice. Before reaching for any tool, it's worth knowing you can find these with something already on your machine.

The quick way: sha256sum

On Linux (GNU coreutils), you can hash every file in a folder and group the ones with identical hashes. Identical hash means identical content, byte for byte.

find . -type f \( -iname "*.jpg" -o -iname "*.jpeg" -o -iname "*.png" \) -exec sha256sum {} \; > hashes.txt
sort hashes.txt | uniq -w64 -D

The first line walks the current folder recursively and writes one hash per image into hashes.txt. The second sorts by hash and prints every line that belongs to a group of identical files. uniq -w64 -D compares only the first 64 characters (the hash itself).

On macOS, use shasum -a 256 instead of sha256sum, and install GNU coreutils (brew install coreutils) to get a uniq that supports -w. It is then available as guniq.

What this gets you, and what it doesn't

This approach is great for exact duplicates: the same file, same bytes, maybe just renamed or moved. It won't catch:

The same photo re-exported at a different resolution
A slightly recompressed JPEG
A screenshot vs. the original image it was taken from

For those "similar but not identical" cases you need perceptual comparison, not a cryptographic hash. That is a different problem and outside the scope of a one-liner.

Also worth noting: the commands above only report duplicates. They don't move or delete anything, which is a good thing. You don't want a script silently deleting files based on a hash match without you checking first.

When you outgrow the one-liner

The commands above work fine for one folder and exact matches. It starts to get annoying once you have:

Thousands of photos across many nested folders
Near-duplicates (same photo, different edit, export or resolution)
A need for a report you can review before anything is touched
A desire to not delete files outright, just quarantine them

That's the situation I built Duplicate Photo Finder Pro for. It is mine and it is paid. It scans folders for exact duplicates (SHA-256, the same method as above) and also for similar images via perceptual hashing with an adjustable threshold. It runs in dry-run mode first, writes a CSV/JSON report, and moves duplicates to a quarantine folder instead of deleting them, keeping whichever copy you choose (oldest, largest or highest resolution) as the original. It runs fully offline, with no cloud upload.

Duplicate Photo Finder Pro, 12 EUR: https://skeilster.gumroad.com/l/nejwxp

Generating Changelogs from Git Log Between Two Tags

Steven Keil — Sun, 25 Oct 2026 12:55:35 +0000

If you've ever had to write release notes by scrolling through git log, you know the drill: squinting at commit messages, manually sorting "fix" from "feat," and copy-pasting into a Markdown file by hand.

Here's a quick way to make that less painful using only git log.

Step 1: Get the raw commits between two tags

git log v1.2.0..v1.3.0 --pretty=format:"%s"

This gives you every commit subject line between v1.2.0 and v1.3.0. You can also use two commit hashes instead of tags if you don't tag releases.

Step 2: Group by conventional commit prefix

If your commits follow a loose convention (feat:, fix:, chore:, docs:), you can split them out with a bit of shell:

git log v1.2.0..v1.3.0 --pretty=format:"%s" | grep "^feat" > features.txt
git log v1.2.0..v1.3.0 --pretty=format:"%s" | grep "^fix"  > fixes.txt
git log v1.2.0..v1.3.0 --pretty=format:"%s" | grep "^chore" > chores.txt

Then concatenate them under headings:

## Features
- feat: add dark mode toggle

## Fixes
- fix: correct timezone offset in export

Step 3: Where this breaks down

This works fine for a handful of commits. It gets tedious once:

Commits don't follow the convention consistently
You have dozens of commits between tags
You want the same grouped Markdown format every release, without rebuilding the grep pipeline each time

That's the exact gap I built a small tool for.

My own product (it is mine and it is paid)

I made a Markdown Changelog Generator that reads the git log between two tags or commits and writes out grouped Markdown automatically, so you don't have to rebuild the grep-and-concatenate pipeline above every release.

It's 12 EUR: https://skeilster.gumroad.com/l/hdsfn

How to clean up merged git branches safely (and the squash-merge trap)

Steven Keil — Sun, 25 Oct 2026 12:55:01 +0000

If you work on a few repos for a year, your local branch list turns into a graveyard: feature/login-fix, tmp-test, hotfix-2, and fifteen more nobody remembers. Cleaning up is easy, but only if you avoid the commands that delete work by accident.

1. See what is already merged

git switch main
git pull
git branch --merged main

Everything in this list has its commits in main already. Deleting these branches loses nothing. Ignore main itself and any branch you are currently on.

2. Delete with the safe flag

git branch -d old-feature

The lowercase -d refuses to delete a branch that is not fully merged. The uppercase -D deletes regardless. Use -d by default and only reach for -D when you have looked at the branch and know you do not need it.

3. Delete many at once, with a preview

git branch --merged main | grep -vE '^\*|main|develop' > to-delete.txt
cat to-delete.txt          # read it first
xargs git branch -d < to-delete.txt

Adjust the grep -v pattern to keep your protected branches. The text file is the preview: nothing is deleted until you have read it.

4. The squash-merge trap

If your team squash-merges pull requests, git branch --merged will not list those branches, because the squashed commit is a new commit. Git cannot tell they are merged. Check the pull request page instead, and only then delete with -D.

5. Remote branches

git fetch --prune
git branch -r --merged origin/main

--prune removes local references to branches that were already deleted on the remote. Deleting remote branches (git push origin --delete name) affects your whole team, so ask first.

6. A habit that avoids most of this

Delete the branch right after the pull request is merged. Most hosting platforms have a setting to do that automatically.

If you do this often

I got tired of repeating these steps across repos and wrote a small Python tool that previews the list and then deletes the merged branches. It is mine and it is paid (9 EUR): https://skeilster.gumroad.com/l/uvpwln. Everything above works without it.

Which edge cases cause you trouble: squash merges, protected branches, something else?

7 things that break a Shopify product CSV import (and how to catch them before you upload)

Steven Keil — Sun, 25 Oct 2026 11:33:25 +0000

If you have ever uploaded a product CSV to Shopify and ended up with a wall of red error messages, or worse, an import that "succeeded" but left you with wrong prices and missing images, this list is for you. These are the problems that cause most failed or silently wrong imports, plus a quick way to check each one yourself.

1. Prices in the wrong format

Spreadsheet tools in many locales write 1.299,00 or 12,5. Shopify expects a plain number with a dot: 1299.00, 12.50. A currency symbol ($12.50) is also a classic.

Check: sort the Variant Price column and look for commas, symbols and letters.

2. Missing prices

A row without Variant Price is either rejected or imported with a price you did not intend.

Check: filter for empty cells in that column.

3. Duplicate SKUs

Two variants with the same Variant SKU confuse inventory tracking and later syncs with other tools.

Check: use a "remove duplicates" or COUNTIF check on the SKU column. Empty SKUs are fine; repeated ones are not.

4. Inventory that is not a whole number

Variant Inventory Qty must be an integer. 3.5, -2 or "ten" will cause trouble.

Check: filter for values containing a dot or a minus sign.

5. Handles with spaces or capitals

The Handle becomes part of the product URL. Use lowercase letters, numbers and dashes: blue-shirt, not Blue Shirt. Rows that belong to the same product (extra variants and images) must repeat the exact same handle.

Check: the same handle should never appear with two different titles.

6. Broken image URLs

Image Src has to be a full, publicly reachable https:// link. A link to a file on your own computer, a missing https://, or a link behind a login does not work.

Check: paste a few URLs into a private browser window.

7. TRUE/FALSE columns that are not TRUE/FALSE

Columns such as Published expect TRUE or FALSE. Values like yes, 1 or true with a trailing space are a common source of "why is my product hidden?".

Check: filter the column and look at the distinct values.

A small habit that saves hours

Before every big import, import three products first and look at them in the admin. Only then upload the full file.

Do it automatically

Doing these seven checks by hand on a file with a few thousand rows is tedious, so I wrote a small checker. It fixes the safe things (spaces, handle format, price format, TRUE/FALSE) and only reports the things that need a human decision (missing prices, duplicate SKUs, bad inventory, broken image URLs). I offer it as a paid service for 15 € per file (up to 2,000 rows), without any login to your store: https://skeilster.gumroad.com/l/shopify-csv-check

Disclosure: that service is mine. The list above works fine without it.

What else has bitten you in a CSV import? I am collecting cases to cover in the checker.