| // main.cpp : Defines the entry point for the console application. | |
| /* | |
| g++ step1-ocr.cpp -I /home/ataosky/software/cvblobs8.3_linux -L /home/ataosky/software/cvblobs8.3_linux -lopencv_core -lopencv_ml -lopencv_imgproc -lopencv_highgui -lblob -o step2-ocr | |
| */ | |
| // | |
| //#include "stdafx.h" //AO | |
| #include "opencv/cv.h" | |
| #include "opencv/highgui.h" | |
| #include "opencv/ml.h" |
#####lshkit的数据格式，load代码
void Matrix<T>::load (std::istream &is)
{
unsigned header[3]; /* entry size, row, col */
assert(sizeof header == 3*4);
is.read((char *)header, sizeof header);
BOOST_VERIFY(is);
BOOST_VERIFY(header[0] == sizeof(T));
reset(header[2], header[1]);
####Scaling Up All Pairs Similarity Search ####论文理解: maxweighti(V)列i最大值 maxweight(x) x[:]的最大值
####代码阅读：
Google的开源实现是All-Pair-Binary的
| /* | |
| This script parse a text file like: | |
| 3 {"list":[[8702,9630,3],[192470,8502,3],[25234,4160,3]]} | |
| into: | |
| 3\t8702\t192470\t25234 | |
| and sort by tagId count acendingly in order to fit to the format described in paper: | |
| Scaling Up All Pairs Similarity Search | |
| to calculate item to item similarity | |
| */ |
| //generate tagId-tagId-cos similarity result | |
| import spark.util.Vector | |
| import scala.math.sqrt | |
| import java.io._ | |
| val word_vec_size=150 | |
| def parseVector(line: String): Vector = { | |
| return new Vector(line.split(' ').slice(1,word_vec_size+1).map(_.toDouble)) | |
| } |
| import scala.util.parsing.json._ | |
| import org.json4s._ | |
| import org.json4s.native.JsonMethods._ | |
| import scala.collection.mutable.ArrayBuffer | |
| import java.io._ | |
| def parseTagTransaction(line:String):ArrayBuffer[String]={ | |
| var tagList = line.split(" ").filter(m => m.length>1); | |
| var result = ArrayBuffer[String]() | |
| if(tagList.length>1) |
Loosely ordered with the commands I use most towards the top. Sublime also offer full documentation.
| Ctrl+C | copy current line (if no selection) |
| Ctrl+X | cut current line (if no selection) |
| Ctrl+⇧+K | delete line |
| Ctrl+↩ | insert line after |
| #!/bin/bash | |
| ## Bash Script to clear cached memory on (Ubuntu/Debian) Linux | |
| ## By Philipp Klaus | |
| ## see <http://blog.philippklaus.de/2011/02/clear-cached-memory-on-ubuntu/> | |
| if [ "$(whoami)" != "root" ] | |
| then | |
| echo "You have to run this script as Superuser!" | |
| exit 1 | |
| fi |
| val sqlContext = new org.apache.spark.sql.hive.HiveContext(sc) | |
| import sqlContext.implicits._ | |
| val pp = sc.makeRDD(1 to 5).map(i => (i, i * 2)).toDF("single", "double") | |
| pp.registerTempTable("people") | |
| sqlContext.sql("select concat('test_',single) from people").collect().foreach(println) |
| sysJupiterDev@gbrdcr00015n02: /bigdata/projects/MERCURY | |
| $ ls spark-1.5.1-bin-hadoop2.6/conf/yarn-conf/ | |
| core-site.xml hadoop-env.sh hdfs-site.xml hive-site.xml mapred-site.xml ssl-client.xml topology.map topology.py yarn-site.xml |