{
    "info": {
        "author": "Behzad Mehrabi",
        "author_email": "bezMehrabi@gmail.com",
        "bugtrack_url": null,
        "classifiers": [],
        "description": "# tojson\nPython HTML document to JSON\n\n#### convert HTML document to JSON\n```\n>>> from tojson import HTML\n>>> with open('sample.html', 'r') as src:\n...    html = HTML(src.read(), text_skip=['html', 'head', 'body'])\n...    html.tojson()) # return json format of html\n{\n  \"html\": {\n    \"head\": {\n      \"title\": {\n        \"text\": \"test\"\n      }\n    },\n    \"body\": {\n      \"bgcolor\": \"FFFFFF\",\n      \"img\": {\n        \"src\": \"clouds.jpg\",\n        \"align\": \"bottom\"\n      },\n      \"a\": [\n        {\n          \"href\": \"http://somegreatsite.com\",\n          \"text\": \"Link Name\"\n        },\n        {\n          \"href\": \"mailto:support@yourcompany.com\",\n          \"text\": \"support@yourcompany.com\"\n        }\n      ],\n      \"h1\": {\n        \"text\": \"This is a Header\"\n      },\n      \"h2\": {\n        \"text\": \"This is a Medium Header\"\n      },\n      \"p\": [\n        {\n          \"text\": \"first paragraph\"\n        },\n        {\n          \"text\": \"second paragraph!\"\n        }\n      ],\n      \"b\": {\n        \"text\": \"This is a new sentence without a paragraph break\",\n        \"i\": {\n          \"text\": \"This is a new sentence without a paragraph break\"\n        }\n      }\n    }\n  }\n}\n```\n#### iterate over tags and their values\n to get tuple contains (tag, value):\n```\n>>> from tojson import HTML\n>>> with open('sample.html', 'r') as src:\n...    html = HTML(src.read(), text_skip=['html', 'head', 'body'])\n...\n>>> for item in html:\n...    item\n...\n('html', {'head': {'title': {'text': 'test'}}, 'body': {'bgcolor': 'FFFFFF', 'img': {'src': 'clouds.jpg', 'align': 'bottom'}, 'a': [{'href': 'http://somegreatsite.com', 'text': 'Link Name'}, {'href': 'mailto:support@yourcompany.com', 'text': 'support@yourcompany.com'}], 'h1': {'text': 'This is a Header'}, 'h2': {'text': 'This is a Medium Header'}, 'p': [{'text': 'first paragraph'}, {'text': 'second paragraph!'}], 'b': {'text': 'This is a new sentence without a paragraph break', 'i': {'text': 'This is a new sentence without a paragraph break'}}}})\n('head', {'title': {'text': 'test'}})\n('title', {'text': 'test'})\n('body', {'bgcolor': 'FFFFFF', 'img': {'src': 'clouds.jpg', 'align': 'bottom'}, 'a': [{'href': 'http://somegreatsite.com', 'text': 'Link Name'}, {'href': 'mailto:support@yourcompany.com', 'text': 'support@yourcompany.com'}], 'h1': {'text': 'This is a Header'}, 'h2': {'text': 'This is a Medium Header'}, 'p': [{'text': 'first paragraph'}, {'text': 'second paragraph!'}], 'b': {'text': 'This is a new sentence without a paragraph break', 'i': {'text': 'This is a new sentence without a paragraph break'}}})\n('img', {'src': 'clouds.jpg', 'align': 'bottom'})\n('a', [{'href': 'http://somegreatsite.com', 'text': 'Link Name'}, {'href': 'mailto:support@yourcompany.com', 'text': 'support@yourcompany.com'}])\n('h1', {'text': 'This is a Header'})\n('h2', {'text': 'This is a Medium Header'})\n('p', [{'text': 'first paragraph'}, {'text': 'second paragraph!'}])\n('b', {'text': 'This is a new sentence without a paragraph break', 'i': {'text': 'This is a new sentence without a paragraph break'}})\n('i', {'text': 'This is a new sentence without a paragraph break'})\n```\n\n\n",
        "description_content_type": "text/markdown",
        "docs_url": null,
        "download_url": "",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "https://github.com/Bezmehrabi/tojson.git",
        "keywords": "",
        "license": "LICENSE.txt",
        "maintainer": "",
        "maintainer_email": "",
        "name": "tojson",
        "package_url": "https://pypi.org/project/tojson/",
        "platform": "",
        "project_url": "https://pypi.org/project/tojson/",
        "project_urls": {
            "Homepage": "https://github.com/Bezmehrabi/tojson.git"
        },
        "release_url": "https://pypi.org/project/tojson/0.1/",
        "requires_dist": null,
        "requires_python": "",
        "summary": "Python Convert HTML Document to JSON",
        "version": "0.1"
    },
    "last_serial": 5583264,
    "releases": {
        "0.1": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "f02ce877485bdc9675258801a5911175",
                    "sha256": "7e0fe52c6fdceab230b2438c997cf0a56b89a5e9a9b529b44f49fb5c4db21fcd"
                },
                "downloads": -1,
                "filename": "tojson-0.1-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "f02ce877485bdc9675258801a5911175",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": null,
                "size": 3736,
                "upload_time": "2019-07-25T13:38:44",
                "url": "https://files.pythonhosted.org/packages/00/2c/8f8941b4280af7cf73c01fbc2869a2190bc7ba5ee95f9a3d3c5f1d56b186/tojson-0.1-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "0ec8e78723eda70e69c739928dd3e50e",
                    "sha256": "ea938c659529322970bc325f371615e92342ce1635901dd87d1384249b090c71"
                },
                "downloads": -1,
                "filename": "tojson-0.1.tar.gz",
                "has_sig": false,
                "md5_digest": "0ec8e78723eda70e69c739928dd3e50e",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 2953,
                "upload_time": "2019-07-25T13:38:47",
                "url": "https://files.pythonhosted.org/packages/2d/85/5e0bc19a718f2844d6b624629411c4972c292cc7b966f90a65b7c8bcf3ae/tojson-0.1.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "f02ce877485bdc9675258801a5911175",
                "sha256": "7e0fe52c6fdceab230b2438c997cf0a56b89a5e9a9b529b44f49fb5c4db21fcd"
            },
            "downloads": -1,
            "filename": "tojson-0.1-py3-none-any.whl",
            "has_sig": false,
            "md5_digest": "f02ce877485bdc9675258801a5911175",
            "packagetype": "bdist_wheel",
            "python_version": "py3",
            "requires_python": null,
            "size": 3736,
            "upload_time": "2019-07-25T13:38:44",
            "url": "https://files.pythonhosted.org/packages/00/2c/8f8941b4280af7cf73c01fbc2869a2190bc7ba5ee95f9a3d3c5f1d56b186/tojson-0.1-py3-none-any.whl"
        },
        {
            "comment_text": "",
            "digests": {
                "md5": "0ec8e78723eda70e69c739928dd3e50e",
                "sha256": "ea938c659529322970bc325f371615e92342ce1635901dd87d1384249b090c71"
            },
            "downloads": -1,
            "filename": "tojson-0.1.tar.gz",
            "has_sig": false,
            "md5_digest": "0ec8e78723eda70e69c739928dd3e50e",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 2953,
            "upload_time": "2019-07-25T13:38:47",
            "url": "https://files.pythonhosted.org/packages/2d/85/5e0bc19a718f2844d6b624629411c4972c292cc7b966f90a65b7c8bcf3ae/tojson-0.1.tar.gz"
        }
    ]
}