AttributeError: 'bytes' object has no attribute 'close' when Tika parser is run

Question

Im trying to run a simple parse line of code using Tika to parse text from a PDF (named outputFileName in this example). This used to run without errors. I recently had my laptop sent in to our work IT for software updates and had to resintall Anaconda and import certain modules (such as Tika 1.22 released on 9 Nov 2019) to run my code. I've also imported the Parser from Tika. Here is the line of code and the error message i get:

Code:

#parsing the text out of pdf file to work with

pdftext = parser.from_file(outputFileName)

Error

    AttributeError                            Traceback (most recent call last)
    <ipython-input-14-239522b5ef31> in <module>
          1 #parsing the text out of pdf file to work with
    ----> 2 pdftext = parser.from_file(outputFileName)

    ~\AppData\Local\Continuum\anaconda3\lib\site-packages\tika\parser.py in from_file(filename, serverEndpoint, xmlContent, headers, config_path, requestOptions)

         34     '''
         35     if not xmlContent:
    ---> 36         jsonOutput = parse1('all', filename, serverEndpoint, headers=headers,config_path=config_path, requestOptions=requestOptions)

         37     else:
         38         jsonOutput = parse1('all', filename, serverEndpoint, services={'meta': '/meta', 'text': '/tika', 'all': '/rmeta/xml'},

    ~\AppData\Local\Continuum\anaconda3\lib\site-packages\tika\tika.py in parse1(option, urlOrPath, serverEndpoint, verbose, tikaServerJar, responseMimeType, services, rawResponse, headers, config_path, requestOptions)
        327     headers.update({'Accept': responseMimeType, 'Content-Disposition': make_content_disposition_header(path)})
        328     status, response = callServer('put', serverEndpoint, service, open(path, 'rb'),
    --> 329                                   headers, verbose, tikaServerJar, config_path=config_path, rawResponse=rawResponse, requestOptions=requestOptions)
        330 
        331     if file_type == 'remote': os.unlink(path)

    ~\AppData\Local\Continuum\anaconda3\lib\site-packages\tika\tika.py in callServer(verb, serverEndpoint, service, data, headers, verbose, tikaServerJar, httpVerbs, classpath, rawResponse, config_path, requestOptions)
        544 
        545     resp = verbFn(serviceUrl, encodedData, **effectiveRequestOptions)
    --> 546     encodedData.close() # closes the file reading data
        547 
        548     if verbose:

    AttributeError: 'bytes' object has no attribute 'close'

I've also downloaded Java 8.

Is there an issue in the tika.py source code? or have i missed somehthing in terms of what needs to be installed? Any help would be much appreciated.

Answer 1

Just install the previous version of tika. I had the same problem and it worked for me. How? Go to a terminal and type pip install tika==1.19