從JavaScript（已編碼）變數中刮除BeautifulSoup

Question

我正在抓取頁面，但無法獲取某個字段，因為它存儲在javascript變量中。

我的問題是，我如何抓取以下代碼，對其進行解碼並保存<li>標簽的內容？ 使用BeautifulSoup和其他任何python。

這是<script>標記內的代碼：

<script type="text/javascript">
    var html_audition_details_sidebar = '    \u003Cdiv id\u003D\u0022apply_wrapper\u0022\u003E        \u003Cdiv class\u003D\u0022header\u0022\u003E            \u003Cp\u003EAudition Information\u003C/p\u003E        \u003C/div\u003E        \u003Cdiv class\u003D\u0022text  \u0022\u003E            \u003Cdiv class\u003D\u0022roleContainer \u0022 style\u003D\u0022color: #999\u003B font\u002Dsize: 14px\u003B\u0022\u003E                \u003Cp\u003EOnly official members can see audition information for this job\u003C/p\u003E            \u003C/div\u003E        \u003C/div\u003E        \u003Cdiv class\u003D\u0022applyButton\u0022\u003E            \u003Cp\u003E\u003Ca class\u003D\u0022applyLink\u0022                                        href\u003D\u0022/accounts/login/apply/41680/\u0022\u003ESubscribe Now                  \u003C/a\u003E\u003C/p\u003E        \u003C/div\u003E    \u003C/div\u003E';
    var html_additional_requirements = '';
    var html_role_listing = '\u003Cdiv class\u003D\u0022text callListing loggedout \u0022\u003E    \u003Cp class\u003D\u0022title\u0022\u003E\u003Ca name\u003D\u0022roles\u0022\u003E\u003C/a\u003ESeeking Talent \u003Cspan class\u003D\u0022optional\u0022\u003ESelect a role below for more information and submission instructions.\u003C/span\u003E\u003C/p\u003E    \u003Cdiv class\u003D\u0022castingRoles\u0022\u003E        \u003Cul\u003E                                \u003Cli \u003E                \u003Ca href\u003D\u0022/casting/untitled\u002Dcomedy\u002Dshort\u002D41680/martinique\u002D159296/\u0022\u003E                    Martinique  (Lead):                \u003Cspan class\u003D\u0022roletag\u0022\u003E                    Female, 18\u002D25, Caucasian                \u003C/span\u003E                \u003Cspan class\u003D\u0022applyNow\u0022\u003E \u003C/span\u003E                \u003C/a\u003E                \u003Cp class\u003D\u0022role\u002Ddesc\u0022                   style\u003D\u0022border\u002Dbottom: none\u003B padding\u002Dbottom: 0px\u003B margin\u002Dbottom: 0px\u003B\u0022\u003E                    native French speaker.                \u003C/p\u003E            \u003C/li\u003E                                            \u003Cli \u003E                \u003Ca href\u003D\u0022/casting/untitled\u002Dcomedy\u002Dshort\u002D41680/justin\u002D159297/\u0022\u003E                    Justin  (Lead):                \u003Cspan class\u003D\u0022roletag\u0022\u003E                    Male, 20\u002D25, All Ethnicities                \u003C/span\u003E                \u003Cspan class\u003D\u0022applyNow\u0022\u003E \u003C/span\u003E                \u003C/a\u003E                \u003Cp class\u003D\u0022role\u002Ddesc\u0022                   style\u003D\u0022border\u002Dbottom: none\u003B padding\u002Dbottom: 0px\u003B margin\u002Dbottom: 0px\u003B\u0022\u003E                    comedy and improv skills, hopeless romantic.                \u003C/p\u003E            \u003C/li\u003E                                            \u003Cli \u003E                \u003Ca href\u003D\u0022/casting/untitled\u002Dcomedy\u002Dshort\u002D41680/flower\u002Dshop\u002Dsalesperson\u002D159299/\u0022\u003E                    Flower Shop Salesperson :                \u003Cspan class\u003D\u0022roletag\u0022\u003E                    Males \u0026amp\u003B Females, 30+, All Ethnicities                \u003C/span\u003E                \u003Cspan class\u003D\u0022applyNow\u0022\u003E \u003C/span\u003E                \u003C/a\u003E                \u003Cp class\u003D\u0022role\u002Ddesc\u0022                   style\u003D\u0022border\u002Dbottom: none\u003B padding\u002Dbottom: 0px\u003B margin\u002Dbottom: 0px\u003B\u0022\u003E                    impatient.                \u003C/p\u003E            \u003C/li\u003E                                            \u003Cli \u003E                \u003Ca href\u003D\u0022/casting/untitled\u002Dcomedy\u002Dshort\u002D41680/models\u002D159300/\u0022\u003E                    Models  (Supporting):                \u003Cspan class\u003D\u0022roletag\u0022\u003E                    Female, 18\u002D35, All Ethnicities                \u003C/span\u003E                \u003Cspan class\u003D\u0022applyNow\u0022\u003E \u003C/span\u003E                \u003C/a\u003E                \u003Cp class\u003D\u0022role\u002Ddesc\u0022                   style\u003D\u0022border\u002Dbottom: none\u003B padding\u002Dbottom: 0px\u003B margin\u002Dbottom: 0px\u003B\u0022\u003E                    small roles, under five lines.                \u003C/p\u003E            \u003C/li\u003E                            \u003C/ul\u003E    \u003C/div\u003E\u003C/div\u003E';
</script>

抱歉，它的格式可怕。

我只需要將所有li標簽內的網址保存為變量html_role_listing 。

例：

轉這個：

\u003Ca href\u003D\u0022/casting/untitled\u002Dcomedy\u002Dshort\u002D41680/martinique\u002D159296/\u0022\u003E

變成這個：

/casting/untitled-comedy-short-41680/martinique-159296/

感謝您的提前幫助！

Answer 1

如果要以通用方式執行此操作，則需要一個解析javascript的庫。 在此示例中，我將使用slimit來做到這一點。

首先，加載您的數據：

from bs4 import BeautifulSoup as Soup
import slimit
from slimit.parser import Parser
from slimit.visitors import nodevisitor    

a = """<script type="text/javascript">
    var html_audition_details_sidebar = '    \u003Cdiv id\u003D\u0022apply_wrapper\u0022\u003E        \u003Cdiv class\u003D\u0022header\u0022\u003E            \u003Cp\u003EAudition Information\u003C/p\u003E        \u003C/div\u003E        \u003Cdiv class\u003D\u0022text  \u0022\u003E            \u003Cdiv class\u003D\u0022roleContainer \u0022 style\u003D\u0022color: #999\u003B font\u002Dsize: 14px\u003B\u0022\u003E                \u003Cp\u003EOnly official members can see audition information for this job\u003C/p\u003E            \u003C/div\u003E        \u003C/div\u003E        \u003Cdiv class\u003D\u0022applyButton\u0022\u003E            \u003Cp\u003E\u003Ca class\u003D\u0022applyLink\u0022                                        href\u003D\u0022/accounts/login/apply/41680/\u0022\u003ESubscribe Now                  \u003C/a\u003E\u003C/p\u003E        \u003C/div\u003E    \u003C/div\u003E';
    var html_additional_requirements = '';
    var html_role_listing = '\u003Cdiv class\u003D\u0022text callListing loggedout \u0022\u003E    \u003Cp class\u003D\u0022title\u0022\u003E\u003Ca name\u003D\u0022roles\u0022\u003E\u003C/a\u003ESeeking Talent \u003Cspan class\u003D\u0022optional\u0022\u003ESelect a role below for more information and submission instructions.\u003C/span\u003E\u003C/p\u003E    \u003Cdiv class\u003D\u0022castingRoles\u0022\u003E        \u003Cul\u003E                                \u003Cli \u003E                \u003Ca href\u003D\u0022/casting/untitled\u002Dcomedy\u002Dshort\u002D41680/martinique\u002D159296/\u0022\u003E                    Martinique  (Lead):                \u003Cspan class\u003D\u0022roletag\u0022\u003E                    Female, 18\u002D25, Caucasian                \u003C/span\u003E                \u003Cspan class\u003D\u0022applyNow\u0022\u003E \u003C/span\u003E                \u003C/a\u003E                \u003Cp class\u003D\u0022role\u002Ddesc\u0022                   style\u003D\u0022border\u002Dbottom: none\u003B padding\u002Dbottom: 0px\u003B margin\u002Dbottom: 0px\u003B\u0022\u003E                    native French speaker.                \u003C/p\u003E            \u003C/li\u003E                                            \u003Cli \u003E                \u003Ca href\u003D\u0022/casting/untitled\u002Dcomedy\u002Dshort\u002D41680/justin\u002D159297/\u0022\u003E                    Justin  (Lead):                \u003Cspan class\u003D\u0022roletag\u0022\u003E                    Male, 20\u002D25, All Ethnicities                \u003C/span\u003E                \u003Cspan class\u003D\u0022applyNow\u0022\u003E \u003C/span\u003E                \u003C/a\u003E                \u003Cp class\u003D\u0022role\u002Ddesc\u0022                   style\u003D\u0022border\u002Dbottom: none\u003B padding\u002Dbottom: 0px\u003B margin\u002Dbottom: 0px\u003B\u0022\u003E                    comedy and improv skills, hopeless romantic.                \u003C/p\u003E            \u003C/li\u003E                                            \u003Cli \u003E                \u003Ca href\u003D\u0022/casting/untitled\u002Dcomedy\u002Dshort\u002D41680/flower\u002Dshop\u002Dsalesperson\u002D159299/\u0022\u003E                    Flower Shop Salesperson :                \u003Cspan class\u003D\u0022roletag\u0022\u003E                    Males \u0026amp\u003B Females, 30+, All Ethnicities                \u003C/span\u003E                \u003Cspan class\u003D\u0022applyNow\u0022\u003E \u003C/span\u003E                \u003C/a\u003E                \u003Cp class\u003D\u0022role\u002Ddesc\u0022                   style\u003D\u0022border\u002Dbottom: none\u003B padding\u002Dbottom: 0px\u003B margin\u002Dbottom: 0px\u003B\u0022\u003E                    impatient.                \u003C/p\u003E            \u003C/li\u003E                                            \u003Cli \u003E                \u003Ca href\u003D\u0022/casting/untitled\u002Dcomedy\u002Dshort\u002D41680/models\u002D159300/\u0022\u003E                    Models  (Supporting):                \u003Cspan class\u003D\u0022roletag\u0022\u003E                    Female, 18\u002D35, All Ethnicities                \u003C/span\u003E                \u003Cspan class\u003D\u0022applyNow\u0022\u003E \u003C/span\u003E                \u003C/a\u003E                \u003Cp class\u003D\u0022role\u002Ddesc\u0022                   style\u003D\u0022border\u002Dbottom: none\u003B padding\u002Dbottom: 0px\u003B margin\u002Dbottom: 0px\u003B\u0022\u003E                    small roles, under five lines.                \u003C/p\u003E            \u003C/li\u003E                            \u003C/ul\u003E    \u003C/div\u003E\u003C/div\u003E';
</script>"""
soup = Soup(a)
js_content = soup.findAll('script')[0].text

現在您可以選擇：使用正則表達式解析js_content或使用詞法分析器（在這種情況下為slimit）。

par = Parser()
tree = par.parse(js_content)
encoded_html = [n.value for n in nodevisitor.visit(tree) if isinstance(n, slimit.ast.String)][2]

最后，您可以輕松地解碼該字符串，它是有效的HTML：

decoded_html = encoded_html.decode('unicode_escape')
print(decoded_html)

因此，請重新解析以下HTML：

role_listing = Soup(decoded_html)
output = [ anchor.attrs['href'] for anchor in role_listing.select('li a') ]
print('---')
print("\n".join(output))

輸出看起來像這樣：

'<div class="text callListing loggedout ">    <p class="title"><a name="roles"></a>Seeking Talent <span class="optional">Select a role below for more information and submission instructions.</span></p>    <div class="castingRoles">        <ul>                                <li >                <a href="/casting/untitled-comedy-short-41680/martinique-159296/">                    Martinique  (Lead):                <span class="roletag">                    Female, 18-25, Caucasian                </span>                <span class="applyNow"> </span>                </a>                <p class="role-desc"                   style="border-bottom: none; padding-bottom: 0px; margin-bottom: 0px;">                    native French speaker.                </p>            </li>                                            <li >                <a href="/casting/untitled-comedy-short-41680/justin-159297/">                    Justin  (Lead):                <span class="roletag">                    Male, 20-25, All Ethnicities                </span>                <span class="applyNow"> </span>                </a>                <p class="role-desc"                   style="border-bottom: none; padding-bottom: 0px; margin-bottom: 0px;">                    comedy and improv skills, hopeless romantic.                </p>            </li>                                            <li >                <a href="/casting/untitled-comedy-short-41680/flower-shop-salesperson-159299/">                    Flower Shop Salesperson :                <span class="roletag">                    Males &amp; Females, 30+, All Ethnicities                </span>                <span class="applyNow"> </span>                </a>                <p class="role-desc"                   style="border-bottom: none; padding-bottom: 0px; margin-bottom: 0px;">                    impatient.                </p>            </li>                                            <li >                <a href="/casting/untitled-comedy-short-41680/models-159300/">                    Models  (Supporting):                <span class="roletag">                    Female, 18-35, All Ethnicities                </span>                <span class="applyNow"> </span>                </a>                <p class="role-desc"                   style="border-bottom: none; padding-bottom: 0px; margin-bottom: 0px;">                    small roles, under five lines.                </p>            </li>                            </ul>    </div></div>'
---
/casting/untitled-comedy-short-41680/martinique-159296/
/casting/untitled-comedy-short-41680/justin-159297/
/casting/untitled-comedy-short-41680/flower-shop-salesperson-159299/
/casting/untitled-comedy-short-41680/models-159300/

Answer 2

無需JS解析器！

我假設您知道您的<script>內容將像示例中一樣進行格式化，並且您已經將該標記的內容script_text到了一個名為script_text的變量中。

首先，我們需要獲取html_role_listing的值，我們可以使用一個好的ol'正則表達式來做到這一點：

>>> import re
>>> html_role_listing_match = re.search(r'var html_role_listing = \'(.+)\';$', script_text, re.MULTILINE)
>>> html_role_listing = html_role_listing_match.group(1)

然后，我們利用\<和類似的轉義序列在Python Unicode字符串中同樣有效（就像它們在JS字符串中一樣）的事實，並使用更安全的eval版本對其進行解析：

>>> import ast
>>> roles_html = ast.literal_eval("u'%s'" % html_role_listing)

只是為了向自己證明這一點，您可以檢查其中的前幾個字符並查看它們是否已正確解析：

>>> print roles_html[:10]
<div class

現在，您有了可以用BeautifulSoup解析的HTML字符串：

>>> from bs4 import BeautifulSoup
>>> soup = BeautifulSoup(roles_html)

並抓住這些鏈接的href屬性。

>>> links = soup.select('li a')
>>> for link in links:
...     print link.attrs['href']
... 
/casting/untitled-comedy-short-41680/martinique-159296/
/casting/untitled-comedy-short-41680/justin-159297/
/casting/untitled-comedy-short-41680/flower-shop-salesperson-159299/
/casting/untitled-comedy-short-41680/models-159300/