beautifulsoup, html5lib: module object has no attribute _base

Question

When I updated my packages I have this new error:

class TreeBuilderForHtml5lib(html5lib.treebuilders._base.TreeBuilder):
AttributeError: 'module' object has no attribute '_base'

I tried to update beautifulsoup, with no more result. How can I fix that?

Nostrum

Answer 1

edit nov, 2017: it seems this doesn't work any more

Finally found out, a search engine didn't throw anything but it's referenced on beautifulsoup's issue tracker: https://bugs.launchpad.net/beautifulsoup/+bug/1603299

it works back with html5lib v0.9999999 (7 nines)

"html5lib<=0.9999999"

Nostrum

This fixes a similar bug in kaggle-cli too – Glutton

(on W7).Unfortunately i tried both downgrading and upgrading.I also tried to seta virtual env using Python 2.7. Nothing worked so far, basically I am stuck on using beautifulsoup library – Authorship

I just overcome this error simply checking that Pycharm was using the wrong interpreter in my virtual env. Indeed in the python idle and Powershell I did not faced such error – Authorship

html5lib<=0.9999999 has a security vulnerability and should not be used any longer. Source: sourceclear.com/registry/security/cross-site-scripting-xss-/… – Intermix

This is the command to fix it: sudo pip install html5lib==0.9999999 – Overthrust

Answer 2

I upgraded beautifulsoup4 and html5lib and it resolved the issue.

pip install --upgrade beautifulsoup4
pip install --upgrade html5lib

Discobolus

Confirmed that this worked to get the sudo pip install spry to work properly without errors February 2017. – Klara

I required --force – Debase

Answer 3

This is an issue with upstream package html5lib: https://bugs.launchpad.net/beautifulsoup/+bug/1603299 To fix, force downgrade to an older version:

pip install --upgrade html5lib==1.0b8

Shook

Answer 4

The downgrade to html5lib 1.0b8 in @Bhavuk answer works but courses a version issue with bleach.

The solution for me was with a change of version of bleach to be compatible with the new version of html5lib

pip install --upgrade bs4
pip install --upgrade bleach==1.4.2
pip install --upgrade html5lib==1.0b8

Python version 3.5

Embrace

For anaconda, I did conda install html5lib==0.9999999, which downgraded bleach to 1.5.0 but it worked – Georganngeorge

Answer 5

The same problem occurred on me. I don't know what you were trying to do, but it occurred on me when I tried to read XML file in pandas, using pd.read_html().

The problem is fixed by upgrading all of beautifulsoup4, html5lib, and lxml, like:

pip install bs4
pip install html5lib
pip install lxml

And restart your Python environment and now it is working.

Foulup

I needed to add -U to upgrade existing installations in some cases, but this worked for me. – Reduplicate

Answer 6

This command solved the problem for me:

 sudo pip install html5lib==0.9999999

Kenweigh

Answer 7

Just install html5lib using this because if you install the normal way then you have to spider using python2.

sudo pip3 install html5lib==0.9999999

Parody

Answer 8

I found trying to switch versions did not work for me. In the end, based on this issue I edited the relevant file at ~/.local/lib/python3.7/site-packages/bs4/builder/_html5lib.py for my purposes.

Helping