phyks
/
bloomysearch
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Read in JS almost done 

TODO : BloomFilters in JS
This commit is contained in:
Phyks 2014-03-02 16:58:52 +01:00
parent d7f9873a35
commit a1fea61f4f
2 changed files with 24 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
index_generation/generate_index.py
View File

@ -25,7 +25,7 @@ def remove_common_words(words):
def padding_16(x):
if x < 256:
return bytes([0,len(samples)])
return bytes([0,x])
else:
return bytes([int(x/256), x%256])
@ -68,10 +68,18 @@ for sample in samples:
"will have to change the way data is stored in the binary "
"file to handle such amount of text.")
#write_little.extend(bitfield(filters[sample].bitarray.length(), 16))
#write_little.extend(filters[sample].bitarray)
#write_big.extend(bitfield(filters[sample].bitarray.length(), 16))
#write_big.extend(filters[sample].bitarray)
tmp = bitarray(endian="little")
tmp.frombytes(padding_16(filters[sample].bitarray.length()))
write_little.extend(tmp)
write_little.extend(filters[sample].bitarray)
write_little.extend([0 for i in range(filters[sample].bitarray.length() %
8)])
tmp = bitarray(endian="big")
tmp.frombytes(padding_16(filters[sample].bitarray.length()))
write_big.extend(tmp)
write_big.extend(filters[sample].bitarray)
write_big.extend([0 for i in range(filters[sample].bitarray.length() %
8)])
with open('../data/search_index_little', 'wb') as index_fh:
print(write_little)

28
js/bloom.js
View File

@ -38,33 +38,27 @@ document.getElementById('search').addEventListener('click', function() {
if (arrayBuffer) {
var tmp = new Uint8Array(arrayBuffer);
var nb_filters = 0;
console.log(tmp);
return;
var nb_filters = tmp[0]*256+tmp[1];
// First 16 bits == number of bitarrays
for (var i = 0; i < 16; i++) {
nb_filters += tmp[i] << i;
}
search_index = new Array(nb_filters);
// For each of the bitarrays, parse it
var offset = 0;
var offset = 2;
for (var i = 0; i < nb_filters; i++) {
// Size of the filter
var length = 0;
for (var j = offset; j < offset + 16; j++) {
length += tmp[j] << j;
}
search_index[i] = new Uint8Array(length);
var length = tmp[offset]*256+tmp[offset+1]; // length is a number of bytes
var length_offset = Math.ceil(length/8);
search_index[i] = new Uint8Array(length_offset);
// Parse filter
for (var j = 16; j < 16 + length; j++) {
search_index[i][j] = tmp[j + offset];
for (var j = 2; j < 2 + length_offset; j++) {
search_index[i][j] = tmp[offset + j];
}
offset += 16 + length;
offset += 2 + length_offset;
}
console.log(search_index);
document.getElementById("loading").innerHTML = "";
loading = false;
usable = true;