Mark-Boger / unpython

Read python pickle files to common lisp

unpython

Read pickled python object structures into Common Lisp.

Only tested on SBCL

Using

CL-USER> (unpython:load-pickle "<path to project>/t/data/character_list.pkl")
("a" "b")
CL-USER> (unpython:load-pickle "<path to project>/t/data/large_dict.pkl")
#<HASH-TABLE :TEST EQUAL :COUNT 21241 {1001A1B873}>
CL-USER> 

Extending

This structure may (will probably) change

To modify how unpython reads in a specific op-code change the perform-op method for that code.

(defmethod perform-op ((op-code (eql +empty-list+)) stream)
    (print "Hello World."))

This will print Hello World every time the +empty-list+ op-code is seen.

There's a wrapper macro to get rid of some of the boiler plate.

(do-for +empty-list+ ()
    (print "Hello World."))

When you give () after the op-code you ignore the stream. If you give (<symbol>) then you can pass that as a handle for the passed stream.

(do-for +empty-list+ (stream)
    (read-byte stream))

Implemented

• Strings
  • Unicode currently has a naive implementation so it is unsafe
• Dictionaries
• Lists

TODO

• Implement other types. (Added as needed)
• Actually do error handling.

License

Do whatever I don't really care.