tobbyvic / cif-parsers

PDBx/mmCIF parsers implemented in Python & JavaScript, available under the MIT license.

cif-parsers

PDBx/mmCIF parsers implemented in Python & JavaScript, available under the MIT license.

The data structure used here is the same as used by the PDBx/mmJSON format, which can be downloaded via PDBj's REST services (https://pdbj.org/help/rest-interface).

Python version usage

import cif
data = cif.__loadCIF__("/path/to/cif/file.cif.gz")

Then, if you want to save the entry as a compressed mmJSON file, you can simply do the following:

import json, gzip
json.dump(data, gzip.open("/path/to/json/file.json.gz", "wb"))

JavaScript version usage

After loading cif.js and loading the data (e.g. via XHR):

var data = loadCIF(request.responseText, 1);

Saving the entry as a compressed mmJSON file is a bit more complex for JavaScript, as this is primarily a client-side language. By using nodejs, it is possible to efficiently convert mmCIF files to mmJSON. The mmjson_manager.js script is used by PDBj to convert the mmCIF files to mmJSON and can serve as an example for those wanting to convert their own CIF files to JSON.

Format

Given the following example CIF/STAR file:

data_PDBID
#
loop_
_pdbx_category1.field1
_pdbx_category1.field2
_pdbx_category1.field3
1 one 1.0
2 two 2.0
3 three 3.0
4 four 4.0
#
loop_
_pdbx_category2.field1
_pdbx_category2.field2
_pdbx_category2.field3
1 one 1.0
2 two 2.0
#
_pdbx_category3.field1  1
_pdbx_category3.field2  one
_pdbx_category3.field3  1.0

After loading the data the data element is in the following format:

{
  "data_PDBID": {
    "pdbx_category1": {
      "field1": [1, 2, 3, 4],
      "field2": ["one", "two", "three", "four"],
      "field3": [1.0, 2.0, 3.0, 4.0]
      
    },
    "pdbx_category2": {
      "field1": [1, 2],
      "field2": ["one", "two"],
      "field3": [1.0, 2.0]
    },
    "pdbx_category3": {
      "field1": [1],
      "field2": ["one"],
      "field3": [1.0]
    }
  }
}

Here, all data objects are elements of an associated array. Each category within a data object is again an element of their parent data object, represented as an associative array. The data items of the category are key-value pairs within this associative array, where the values are an array of data values to include loop_ structures. No discrimination is made between loop_ and regular structures, both are stored as an associative array. The "." and "?" CIF values are both saved as null values in mmJSON, since undefined is unfortunately not available in JSON in order to distinguish the two.