Reading of number of samples working now in both JS and Python

This commit is contained in:
Phyks 2014-02-26 22:48:15 +01:00
parent 221eeaf8f5
commit d7f9873a35
5 changed files with 45 additions and 154 deletions

2
.gitignore vendored
View File

@ -1,3 +1,3 @@
*/__pycache__
*/search_index
*/search_index*
*/pages_index.json

View File

@ -13,7 +13,6 @@
</p>
</form>
<div id="results"></div>
<script type="text/javascript" src="js/bloomfilter.js"></script>
<script type="text/javascript" src="js/bloom.js"></script>
</body>
</html>

View File

@ -23,15 +23,21 @@ def remove_common_words(words):
returned = [word for word in words if len(word) > 3]
return returned
def bitfield(n, fill):
return [1 if digit=='1' else 0 for digit in bin(n)[2:].zfill(fill)]
def padding_16(x):
if x < 256:
return bytes([0,len(samples)])
else:
return bytes([int(x/256), x%256])
# =============================================================================
samples = list_directory("../samples/")
filters = {}
p = stemmer.PorterStemmer()
write = bitarray(bitfield(len(samples), 16))
write_little = bitarray(endian="little")
write_big = bitarray(endian="big")
write_little.frombytes(padding_16(len(samples)))
write_big.frombytes(padding_16(len(samples)))
if len(samples) > 65535:
sys.exit("[ERROR] Too many articles to index. You will have to change the "
@ -62,11 +68,17 @@ for sample in samples:
"will have to change the way data is stored in the binary "
"file to handle such amount of text.")
write.extend(bitfield(filters[sample].bitarray.length(), 16))
write.extend(filters[sample].bitarray)
#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)
with open('../data/search_index', 'wb') as index_fh:
index_fh.write(write.tobytes())
with open('../data/search_index_little', 'wb') as index_fh:
print(write_little)
write_little.tofile(index_fh)
with open('../data/search_index_big', 'wb') as index_fh:
print(write_big)
write_big.tofile(index_fh)
with open('../data/pages_index.json', 'w') as pages_fh:
pages_fh.write(json.dumps(samples))

View File

@ -2,8 +2,19 @@ var loading = false;
var usable = false;
var search_index = false;
window.onload = function() {
};
// Check endianness to serve right file
function checkEndian(){
var a = new ArrayBuffer(4);
var b = new Uint8Array(a);
var c = new Uint32Array(a);
b[0] = 0xa1;
b[1] = 0xb2;
b[2] = 0xc3;
b[3] = 0xd4;
if(c[0] == 0xd4c3b2a1) return "little";
if(c[0] == 0xa1b2c3d4) return "big";
else return 0;
}
document.getElementById('search_form').addEventListener('submit', function(e) {
e.preventDefault();
@ -19,18 +30,17 @@ document.getElementById('search').addEventListener('click', function() {
document.getElementById("loading").innerHTML = "Loading index file...";
var oReq = new XMLHttpRequest();
oReq.open("GET", "/data/search_index", true);
oReq.open("GET", "data/search_index_"+checkEndian(), true);
oReq.responseType = "arraybuffer";
oReq.onload = function (oEvent) {
var arrayBuffer = oReq.response; // Note: not oReq.responseText
if (arrayBuffer) {
loading = false;
usable = true;
document.getElementById("loading").innerHTML = "";
if (arrayBuffer) {
var tmp = new Uint8Array(arrayBuffer);
var nb_filters = 0;
console.log(tmp);
return;
// First 16 bits == number of bitarrays
for (var i = 0; i < 16; i++) {
@ -55,6 +65,9 @@ document.getElementById('search').addEventListener('click', function() {
offset += 16 + length;
}
document.getElementById("loading").innerHTML = "";
loading = false;
usable = true;
}
else {
document.getElementById("loading").innerHTML = "Error while loading search index.";
@ -78,20 +91,20 @@ document.getElementById('search').addEventListener('click', function() {
}
});
function callback_change() {
/*function callback_change() {
if(!usable) {
return;
}
var search = document.getElementById("search").value;
document.getElementById("results").innerHTML = "<h2>Results :</h2>";
/*TODO for(var key in index) {
//* for(var key in index) {
if(index[key].test(search)) {
document.getElementById("results").innerHTML += "<p>"+key+"</p>";
}
}*/
}* //
if(!document.querySelectorAll("#results p").length) {
document.getElementById("results").innerHTML += "<p>No results...</p>";
}
}
document.getElementById("search").addEventListener('input', callback_change);
document.getElementById("search").addEventListener('input', callback_change);*/

View File

@ -1,133 +0,0 @@
(function(exports) {
exports.BloomFilter = BloomFilter;
exports.fnv_1a = fnv_1a;
exports.fnv_1a_b = fnv_1a_b;
var typedArrays = typeof ArrayBuffer !== "undefined";
// Creates a new bloom filter. If *m* is an array-like object, with a length
// property, then the bloom filter is loaded with data from the array, where
// each element is a 32-bit integer. Otherwise, *m* should specify the
// number of bits. *k* specifies the number of hashing functions.
function BloomFilter(m, k) {
var a;
if (typeof m !== "number") a = m, m = a.length * 32;
this.m = m;
this.k = k;
var n = Math.ceil(m / 32),
i = -1;
if (typedArrays) {
var kbytes = 1 << Math.ceil(Math.log(Math.ceil(Math.log(m) / Math.LN2 / 8)) / Math.LN2),
array = kbytes === 1 ? Uint8Array : kbytes === 2 ? Uint16Array : Uint32Array,
kbuffer = new ArrayBuffer(kbytes * k),
buckets = this.buckets = new Int32Array(n);
if (a) while (++i < n) buckets[i] = a[i];
this._locations = new array(kbuffer);
} else {
var buckets = this.buckets = [];
if (a) while (++i < n) buckets[i] = a[i];
else while (++i < n) buckets[i] = 0;
this._locations = [];
}
}
// See http://willwhim.wordpress.com/2011/09/03/producing-n-hash-functions-by-hashing-only-once/
BloomFilter.prototype.locations = function(v) {
var k = this.k,
m = this.m,
r = this._locations,
a = fnv_1a(v),
b = fnv_1a_b(a),
i = -1,
x = a % m;
while (++i < k) {
r[i] = x < 0 ? (x + m) : x;
x = (x + b) % m;
}
return r;
};
BloomFilter.prototype.add = function(v) {
var l = this.locations(v + ""),
i = -1,
k = this.k,
buckets = this.buckets;
while (++i < k) buckets[Math.floor(l[i] / 32)] |= 1 << (l[i] % 32);
};
BloomFilter.prototype.test = function(v) {
var l = this.locations(v + ""),
i = -1,
k = this.k,
b,
buckets = this.buckets;
while (++i < k) {
b = l[i];
if ((buckets[Math.floor(b / 32)] & (1 << (b % 32))) === 0) {
return false;
}
}
return true;
};
// Estimated cardinality.
BloomFilter.prototype.size = function() {
var buckets = this.buckets,
bits = 0;
for (var i = 0, n = buckets.length; i < n; ++i) bits += popcnt(buckets[i]);
return -this.m * Math.log(1 - bits / this.m) / this.k;
};
// http://graphics.stanford.edu/~seander/bithacks.html#CountBitsSetParallel
function popcnt(v) {
v -= (v >> 1) & 0x55555555;
v = (v & 0x33333333) + ((v >> 2) & 0x33333333);
return ((v + (v >> 4) & 0xF0F0F0F) * 0x1010101) >> 24;
}
// Fowler/Noll/Vo hashing.
function fnv_1a(v) {
var n = v.length,
a = 2166136261,
c,
d,
i = -1;
while (++i < n) {
c = v.charCodeAt(i);
if (d = c & 0xff000000) {
a ^= d >> 24;
a += (a << 1) + (a << 4) + (a << 7) + (a << 8) + (a << 24);
}
if (d = c & 0xff0000) {
a ^= d >> 16;
a += (a << 1) + (a << 4) + (a << 7) + (a << 8) + (a << 24);
}
if (d = c & 0xff00) {
a ^= d >> 8;
a += (a << 1) + (a << 4) + (a << 7) + (a << 8) + (a << 24);
}
a ^= c & 0xff;
a += (a << 1) + (a << 4) + (a << 7) + (a << 8) + (a << 24);
}
// From http://home.comcast.net/~bretm/hash/6.html
a += a << 13;
a ^= a >> 7;
a += a << 3;
a ^= a >> 17;
a += a << 5;
return a & 0xffffffff;
}
// One additional iteration of FNV, given a hash.
function fnv_1a_b(a) {
a += (a << 1) + (a << 4) + (a << 7) + (a << 8) + (a << 24);
a += a << 13;
a ^= a >> 7;
a += a << 3;
a ^= a >> 17;
a += a << 5;
return a & 0xffffffff;
}
})(typeof exports !== "undefined" ? exports : this);