Determine file type of a string

Question

PHP ships with various methods of identifying the type of a file, but is it possible to identify a data type when the file in question only exists as a binary string representation and not as an actual file on disc?

The reason for this is I'm doing some maintenance work on a CMS where the previous developer, being a bit of a wally, decided to store image data into the system as database BLOBs. My current project is dumping the BLOBs out into files, and saving the path to the files into the database in place of the BLOBs.

As I said, my predecessor was a bit of a wally and not only did he store all this data as BLOBs, he also didn't save the datatype of the data anywhere.

The migration utility I wrote for part of this project saves the file to disc without an extension, tries to determine the type of the file with exif_imagetype() and if it manages to identify the file type, renames the file with the correct extension.

However, the classes that use the image data also need updating so they can continue to function with paths and files on disc instead of BLOBs.

The methods that create and update images expect binary strings (to BLOB into the database) and in an ideal world I'd rather rewrite these methods to use is_uploaded_file, move_uploaded_file, etc. However, there's no evidence anywhere in the class of direct manipulation of the $_FILES array so the filedata obviously comes from outside the classbut given how convoluted the code is (and no comments to help out) I can't find it.

As a stopgap solution until I finally track down the actual file upload management code, I plan to manipulate the file data as strings in the class as is currently done, but saving the strings to files instead of into the database. This should minimize the impact on other parts of the codebase that are relying on this class.

I could just do what the migration script is doing and rename the file after saving and then identifying it, but this could prove problematic in the case where there is already a file there. I'd rather know what the data type is before I commit the data to disc.

Answer 1

finfo_buffer() is what you want. You can pass it your string and it will tell you what the file type is based on your mime.magic file.

More info here: https://www.php.net/manual/en/function.finfo-buffer.php

Answer 2

If you are working only with images, you can find the filetype looking in the first few bytes of the string - PNG files begin with 0x89 0x50 0x4E 0x47 0x0D 0x0A 0x1A 0x0A, GIF - with GIF, and JPEG contains FFD8 in the header. I've wrote script for parsing headers of those 3 types, but since I don't have it here, I'll update my answer as soon as I get it

Answer 3

You can use finfo_buffer. It works on strings rather than on-disk files.