From 7f2f7f7c3e5d4c626d7e8d4d45fee2b11b6d5221 Mon Sep 17 00:00:00 2001
From: Matt Strapp <strap012@umn.edu>
Date: Thu, 24 Sep 2020 07:24:19 -0500
Subject: rearrange

---
 P1/include/mapper.h    | 46 ++++++++++++++++++++++++++++++++++++++++++++
 P1/include/mapreduce.h |  6 ++++++
 P1/include/reducer.h   | 24 +++++++++++++++++++++++
 P1/include/utils.h     | 52 ++++++++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 128 insertions(+)
 create mode 100644 P1/include/mapper.h
 create mode 100644 P1/include/mapreduce.h
 create mode 100644 P1/include/reducer.h
 create mode 100644 P1/include/utils.h

(limited to 'P1/include')

diff --git a/P1/include/mapper.h b/P1/include/mapper.h
new file mode 100644
index 0000000..2e3693c
--- /dev/null
+++ b/P1/include/mapper.h
@@ -0,0 +1,46 @@
+#ifndef MAPPER_H
+#define MAPPER_H
+
+#include "utils.h"
+
+// ###### DO NOT REMOVE ######
+#define MAXKEYSZ 100
+#define MAXVALUESZ 100
+
+// ###### DO NOT REMOVE ######
+char *mapOutDir;
+int mapperID;
+
+
+// You are free to change the intermediate data structure as it suits you
+// If you do so, ensure the provided utility functions are also changed
+// 1 1 1...
+typedef struct valueList {
+	// MAXVALUESZ can be reduced to a small value as you are only storing "1"
+	char value[MAXVALUESZ];
+	struct valueList *next;
+}valueList;
+
+// word 1 1 1...
+typedef struct intermediateDS{
+	char key[MAXKEYSZ];
+	valueList *value;
+	struct intermediateDS *next;
+}intermediateDS;
+
+// ###### DO NOT REMOVE ######
+valueList *createNewValueListNode(char *value);
+valueList *insertNewValueToList(valueList *root, char *count);
+void freeValueList(valueList *root);
+
+// ###### DO NOT REMOVE ######
+intermediateDS *createNewInterDSNode(char *word, char *count);
+intermediateDS *insertPairToInterDS(intermediateDS *root, char *word, char *count);
+void freeInterDS(intermediateDS *root);
+
+// ###### DO NOT REMOVE ######
+void emit(char *key, char *value);
+void map(char *chunkData);
+void writeIntermediateDS();
+
+#endif
\ No newline at end of file
diff --git a/P1/include/mapreduce.h b/P1/include/mapreduce.h
new file mode 100644
index 0000000..9b5950b
--- /dev/null
+++ b/P1/include/mapreduce.h
@@ -0,0 +1,6 @@
+#ifndef MAPREDUCE_H
+#define MAPREDUCE_H
+
+#include "utils.h" //sendChunkData and shuffle
+
+#endif
\ No newline at end of file
diff --git a/P1/include/reducer.h b/P1/include/reducer.h
new file mode 100644
index 0000000..79ea452
--- /dev/null
+++ b/P1/include/reducer.h
@@ -0,0 +1,24 @@
+#ifndef REDUCER_H
+#define REDUCER_H
+
+#include "utils.h"
+
+#define MAXKEYSZ 50
+
+// ###### DO NOT REMOVE ######
+typedef struct finalKeyValueDS {
+	char key[MAXKEYSZ];
+	int value;
+	struct finalKeyValueDS *next;
+} finalKeyValueDS;
+
+// ###### DO NOT REMOVE ######
+finalKeyValueDS *createFinalKeyValueNode(char *value, int count);
+finalKeyValueDS *insertNewKeyValue(finalKeyValueDS *root, char *word, int count);
+void freeFinalDS(finalKeyValueDS *root);
+
+// ###### DO NOT REMOVE ######
+void writeFinalDS(int reducerID);
+void reduce(char *key);
+
+#endif
\ No newline at end of file
diff --git a/P1/include/utils.h b/P1/include/utils.h
new file mode 100644
index 0000000..cfe56f2
--- /dev/null
+++ b/P1/include/utils.h
@@ -0,0 +1,52 @@
+#ifndef UTILS_H
+#define UTILS_H
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <errno.h>
+#include <unistd.h>
+#include <dirent.h>
+#include <string.h>
+#include <fcntl.h>
+#include <sys/msg.h>
+#include <sys/stat.h>
+#include <sys/wait.h>
+#include <sys/types.h>
+#include <ctype.h>
+
+#define chunkSize 1024
+#define MSGSIZE 1100
+#define ENDTYPE 1000
+#define ACKTYPE 1100
+
+struct msgBuffer {
+    long msgType;
+    char msgText[MSGSIZE];
+};
+
+// mapper side
+int validChar(char c);
+// getWord usage - retrieves words from the chunk passed until it is fully traversed
+// given a chunk of data chunkData, the call to getWord should look as below:
+// int i = 0;
+// char *buffer;
+// while ((buffer = getWord(chunkData, &i)) != NULL){
+// 			your code
+// }
+char *getWord(char *chunk, int *i);
+char *getChunkData(int mapperID);
+void sendChunkData(char *inputFile, int nMappers);
+
+
+// reducer side
+int hashFunction(char* key, int reducers);
+int getInterData(char *key, int reducerID);
+void shuffle(int nMappers, int nReducers);
+
+// directory
+void createOutputDir();
+char *createMapDir(int mapperID);
+void removeOutputDir();
+void bookeepingCode();
+
+#endif
\ No newline at end of file
-- 
cgit v1.2.3