etherpump/etherdump/commands/index.py

#!/usr/bin/env python

from __future__ import print_function
from argparse import ArgumentParser
import json, os, re
from urllib import urlencode
from urllib2 import urlopen, HTTPError, URLError
from jinja2 import FileSystemLoader, Environment
from datetime import datetime


def group (items, key=lambda x: x):
    ret = []
    keys = {}
    for item in items:
        k = key(item)
        if k not in keys:
            keys[k] = []
        keys[k].append(item)
    for k in sorted(keys):
        keys[k].sort()
        ret.append(keys[k])
    return ret

def main(args):
    p = ArgumentParser("")
    p.add_argument("input", nargs="+", help="filenames")
    p.add_argument("--templates", default=None, help="templates path")
    args = p.parse_args(args)

    tmpath = args.templates
    if tmpath == None:
        tmpath = os.path.split(os.path.abspath(__file__))[0]
        tmpath = os.path.split(tmpath)[0]
        tmpath = os.path.join(tmpath, "data", "templates")

    env = Environment(loader=FileSystemLoader(tmpath))
    template = env.get_template("index.html")

    def base (x):
        return re.sub(r"(\.raw\.html)|(\.diff\.html)|(\.meta\.json)|(\.raw\.txt)$", "", x)

    inputs = args.input
    inputs.sort()
    inputs = group(inputs, base)

    def loadmeta(paths):
        for p in paths:
            if p.endswith(".meta.json"):
                with open(p) as f:
                    return json.load(f)

    inputs = map(loadmeta, inputs)
    # sort by last edited (reverse)
    timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
    print (template.render({"timestamp": timestamp, "pads": inputs}).encode("utf-8"))

    # TODO: MODIFY THIS TO MAKE THE OUTPUT JOINABLE with the collected META DATA
    # evt: how can the metadata become a GRAPH structure!!! with each output DOCUMENT
    # 
    # print ("<ol>")
    # for x in inputs:
    #     padid = x
    #     metapath = os.path.join(x, "{0}.meta.json".format(padid))
    #     if os.path.exists(metapath):
    #         print ("""<li><a href="{0}">{0}</a></li>""".format(x))
    #         with open(metapath) as f:
    #             meta = json.load(f)
    #         indexpath = os.path.join(x, "index.html")
    #         with open(indexpath, "w") as f:

    # print ("</ol>")