# `recover-pdfs` recovers all PDFs from any kind of disk or filesystem.

- Did you accidentally delete one or all of your PDF documents?
- Do you need an important PDF bank statement from an old computer, but the 2008 hard-drive doesn't boot?

I built `recover-pdf` as a filesystem-agnostic tool for recovering PDF documents.
It literally looks over every bit of information on your disk and anything that looks
like a PDF gets put into a recovery directory.

Usage: `recover-pdfs <DISK_OR_IMAGE_FILE> <BACKUP_DIRECTORY`

DISK_OR_IMAGE_FILE can be replaced with the actual disk device in /dev, e.g. /dev/sda1.
Alternatively, DISK_OR_IMAGE_FILE can be an ISO or any similar image file.  Basically
it can be anything made up of 1s and 0s and readable by C file APIs.

BACKUP_DIRECTORY iss the directory/folder where you'd like the
backed up PDFs to go.  The program does not do de-duplication.
This can be done with other tools, such as `duff *`.

# Build, install, and run

```{bash}
# Compile the program
make
# Optionally, install it to /usr/bin
# sudo cp recover-pdfs /usr/bin
# Set up a backup directory
mkdir -p recovered_pdfs
# Run the program
./recover-pdfs old_laptop.img recovered_pdfs
```

# Alternative workflow with pipe viewer

If you'd like a progress bar, you can use pipe viewer to see how much of the
disk/image file has been processed.

```{bash}
# Install pipe viewer if you don't have it already
sudo apt install pv
# Run the program with pipe viewer to see progress
pv old_laptop.img | ./recover-pdfs - recovered_pdfs
```

# Recommended followup routine

```{bash}
# De-duplicate recovered PDFs
(cd recovered_pdfs && rm $(duff -e *))
# Sort by size to quickly identify 'run-awy' PDFs that are hundreds of megabytes or more
du -sh $(du -sb recovered_pdfs | sort -n) 2>/dev/null
```

By default, only PDFs up to 100MB are recovered.  If you have a lot of large
PDFs, you can increase this limit at the top of the `recover-pdfs.c` source file, and then recompile.

Specifically, change this line:
```{c}
#define MAX_PDF_SIZE (100LL * 1024 * 1024)
```