fgets function returning values as python string?

[Francisco Lahuerta Calahorra]

HI I trying to build a function that read IO a file and finds the position of a Marker

I would like to do it in cython with the C functions fopen and fgets

but I do not know how to convert the tring that I get in the fgets into a python string.

In the following example a file is read as python and with the c functions, when is 

read as python I find the marker but when a read with the C function I can not find the marker.

Would like to ask how can be done so?

    def _read_seq1 (self, filename, MARKER, n_frame, BLOCK_SIZE):        
       
        filename_byte_string = filename.encode("UTF-8")
       
        cdef char* fname = filename_byte_string
        cdef FILE* cfl
        cfl = fopen(fname, "rb")
        fl = open(filename, 'rb')
       
        if cfl == NULL:
            return -1, ""
           
        cdef size_t _block_size = BLOCK_SIZE
        string_val = "x" * (BLOCK_SIZE + 1)
        cdef char* block = string_val
       
        current = ""
        frame_data = ""
        ii_frame = 0
        count = 0
       
        current1 = ""
        block1 = ""
       
        while True:
           
            if fgets(block, _block_size, cfl) == NULL: break              
           
            block1 = fl.read(BLOCK_SIZE)
            current1 += block1
           
            current += unicode(block, "ISO-8859-1").encode('utf8') ### How to do this part
           
            markerpos = current.find( MARKER )
            markerpos1 = current1.find( MARKER )
           
            if markerpos >= 0: print "Cython", markerpos
            if markerpos1 >= 0: print "Python", markerpos1, " ", len( current1)
               
            count = count + 1
            if count == 5: break
           
        #fl.close()  
        fclose(cfl)
        fl.close()
       
        return ii_frame, frame_data

 

[Nils Bruin]

On Tuesday, February 10, 2015 at 2:25:37 AM UTC-8, Francisco Lahuerta Calahorra wrote:

           
            block1 = fl.read(BLOCK_SIZE)
            current1 += block1
           
            current += unicode(block, "ISO-8859-1").encode('utf8') ### How to do this part
           

It looks like you want to *decode* an ISO-8859-1 stream into a python unicode object. Probably PyUnicode_DecodeLatin1 (CPython C-API) does this. Note that "current +=" will create a new object every time, so your complexity is O(<length>^2) rather than the O(<length>) it should be.

You can then of course *encode* your unicode string into a utf8 byte-sequence once you've done that with ".encode('utf8'), but that is not what fl.read does.

[Francisco Lahuerta Calahorra]

Yes is what fl.read does, but I want to avoid the use of the python fl.read and use fgets instead, that is a C++ function.

However my problem is that the text that I read with fgets is not encoded as the text that I read with fl.read.

and I would like to know how to read with fgets and encode that text as when is read with  fl.read.

[Nils Bruin]

On Wednesday, February 11, 2015 at 4:43:11 AM UTC-8, Francisco Lahuerta Calahorra wrote:

Yes is what fl.read does, but I want to avoid the use of the python fl.read and use fgets instead, that is a C++ function.

However my problem is that the text that I read with fgets is not encoded as the text that I read with fl.read.

and I would like to know how to read with fgets and encode that text as when is read with  fl.read.

I think you are using the word "encode" for what the python documentation calls "decode" . It might be a good idea to read the python reference section about unicode, and in particular the C-API documentation of PyUnicode_DecodeLatin1, which is almost certainly what you need.

(also: don't top-post).