Merge remote-tracking branch 'origin/indexer' into indexer

Makefile

+# the compiler: gcc for C program, define as g++ for C++
+  CC = g++
+
+  # compiler flags:
+  #  -g    adds debugging information to the executable file
+  #  -Wall turns on most, but not all, compiler warnings
+  CFLAGS  = -std=c++11 -c
+
+  # the build target executable:
+  TARGET = indexer/tests/test
+
+  all: $(TARGET)
+
+  $(TARGET): $(TARGET).cpp
+		$(CC) $(CFLAGS) $(TARGET).cpp -o $(TARGET)
+  clean:
+		$(RM) $(TARGET)
+
+test:
+	make all
+	    chmod a+x indexer/tests/test
+		./indexer/tests/test indexer/tests/source/test1.txt > indexer/tests/source/test1out.txt
+		diff test1out.txt test1.out.correct.txt
+

indexer/Indexer.cpp

@@ -27,6 +27,18 @@ void Indexer::run() {
     save();
 }
 
+void Indexer::verbose_run() {
+    while(pointerToDictionaries.Size() != 0) {
+        unordered_map<string, vector<int>> dictionary = *pointerToDictionaries.Pop();
+        for(auto word : dictionary) {
+	        for(auto location : word.second) {
+                indexedCount++;
+                masterDictionary[word.first].push_back(location);
+                }
+            }
+        }
+    }
+
 void Indexer::save() {
     map<string, vector<size_t> > maps(masterDictionary.begin(), masterDictionary.end());
     string fileName = "index" + to_string(currentFile) + ".txt";
@@ -42,7 +54,19 @@ void Indexer::save() {
     }
     close(file);
     currentFile++;
-}
+    }
+
+void Indexer::verbose_save() {
+    map<string, vector<size_t> > maps(masterDictionary.begin(), masterDictionary.end());
+    for(auto word : maps) {
+        cout << word.first << endl;
+        for(auto location : word.second) {
+            cout << location << " ";
+            }
+	    cout << endl;
+        }
+    currentFile++;
+    }
 
 void Indexer::reset() {
     masterDictionary.clear();

indexer/Indexer.h

+#ifndef indexer_h
+#define indexer_h
 #include "../ProducerConsumerQueue.h"
 #include "../ProducerConsumerQueue.cpp"
 #include <unordered_map>
@@ -8,6 +10,7 @@
 #include <fcntl.h>
 #include <unistd.h>
 
+
 /*
 
 Objective: Pulls small dictionaries from the parser and merges them into the 
@@ -20,8 +23,10 @@ using namespace std;
 class Indexer {
     public:
         Indexer();
-        void run();
-        ProducerConsumerQueue<unordered_map<string, vector<int> >*> pointerToDictionaries;
+		void run();
+		void verbose_run();
+		void verbose_save();
+		ProducerConsumerQueue<unordered_map<string, vector<int> > * > pointerToDictionaries;
     private:
         void save();
         void reset();
@@ -30,4 +35,6 @@ class Indexer {
         size_t currentFile;
         size_t totalIndexed;
         size_t currentlyIndexed;
-};
\ No newline at end of file
+};
+
+#endif /*indexer_h*/

indexer/tests/source/test1.out.correct.txt

+and
+34 36
+anticipated
+13
+are
+15 31
+back
+17
+been
+39
+better
+11
+companies
+14
+country
+20
+cut
+41
+economy
+2
+even
+10
+good
+6
+great
+33
+have
+38
+in
+7
+into
+18
+is
+3
+jobs
+42 43 44
+leaving
+27
+long
+23
+looking
+4 32
+massively
+40
+my
+8
+numbers
+30
+of
+26
+opinion
+9
+our
+19
+pouring
+16
+regulations
+35
+reversing
+21
+taxes
+37
+term
+24
+than
+12
+the
+0 22 28
+trend
+25
+unemployment
+29
+us
+1
+very
+5
\ No newline at end of file

indexer/tests/source/test1.txt

+The U.S. economy is looking very good, in my opinion,
+even better than anticipated. Companies are pouring back into our country,
+reversing the long term trend of leaving. The unemployment numbers are looking great,
+and Regulations and Taxes have been massively Cut! JOBS, JOBS, JOBS
\ No newline at end of file

indexer/tests/test.cpp

+//
+// Created by Zane Dunnings on 2/19/18.
+//
+
+#include <iostream>
+#include <fstream>
+#include <string>
+#include <algorithm>
+#include "../Indexer.h"
+
+using namespace std;
+
+int main( int argc, char *argv[]) {
+	Indexer indexer = Indexer();
+
+	unordered_map<string, vector<int> > test1;
+	ifstream ifstream1(argv[ 1 ]);
+
+	string word = "";
+	int id = 0;
+	while(ifstream1 >> word) {
+		std::transform(word.begin(), word.end(), word.begin(), ::tolower);
+		word.erase(remove_if(word.begin(), word.end(), [](char c) { return !isalpha(c); } ), word.end());
+		if(word != "")
+			test1[word].push_back(id);
+		id++;
+		}
+
+	indexer.pointerToDictionaries.Push(&test1);
+	indexer.verbose_run();
+	indexer.verbose_save();
+	cout << "congrats!";
+
+	}
\ No newline at end of file