Add a license and details in the data collected dire.

frmichel · Sep 7, 2013 · 1bc3b26 · 1bc3b26
1 parent 6044874
commit 1bc3b26
Show file tree

Hide file tree

Showing 4 changed files with 64 additions and 50 deletions.
diff --git a/LICENSE.txt b/LICENSE.txt
@@ -1,3 +1,5 @@
+--MIT Lience --
+
 Copyright (C) 2013 by Franck MICHEL
 
 Permission is hereby granted, free of charge, to any person obtaining a copy

diff --git a/src/fm/flickr/api/wrapper/service/FlickrService.java b/src/fm/flickr/api/wrapper/service/FlickrService.java
@@ -802,46 +802,49 @@ private Map<String, Long> findDateInGroupByDicotomy(String groupId, Long searche
 			// Get the first (morst recent) and last (oldest) photos of the page
 			NodeList photosList = xmlResp.getElementsByTagName("photo");
 			int nbPhotos = photosList.getLength();
-			Element firstphoto = (Element) photosList.item(0);
-			Element lastphoto = (Element) photosList.item(nbPhotos - 1);
-
-			// If the searched date is before the last (oldest) photo, ie.
-			// if date < date(last photo), then we must look further in the past
-			if (searchedDate < Long.valueOf(lastphoto.getAttribute("dateadded"))) {
-				if (page + increment < maxPages)
-					page += increment;
-				logger.trace("searchedDate < " + lastphoto.getAttribute("dateadded") + ": going further in the past");
-			}
+			if (nbPhotos > 0) {
 
-			// If the searched date is between the first (most recent) and the last (oldest) photos of the page, ie.
-			// older than the first photo and more recent than the last photo, ie.
-			// if date(last photo) <= searchedDate <= date(first photo), then we have found the right page!
-			else if (Long.valueOf(lastphoto.getAttribute("dateadded")) <= searchedDate && searchedDate <= Long.valueOf(firstphoto.getAttribute("dateadded"))) {
-				indexInPage = 0;
-				while (indexInPage < photosList.getLength()) {
-					Element photo = (Element) photosList.item((int) indexInPage);
-					if (searchedDate > Long.valueOf(photo.getAttribute("dateadded")))
-						break;
-					indexInPage++;
-					// When this inner loop is over, we have found the first photo, in the page, that is more recent than startDate
+				Element firstphoto = (Element) photosList.item(0);
+				Element lastphoto = (Element) photosList.item(nbPhotos - 1);
+
+				// If the searched date is before the last (oldest) photo, ie.
+				// if date < date(last photo), then we must look further in the past
+				if (searchedDate < Long.valueOf(lastphoto.getAttribute("dateadded"))) {
+					if (page + increment < maxPages)
+						page += increment;
+					logger.trace("searchedDate < " + lastphoto.getAttribute("dateadded") + ": going further in the past");
 				}
-				foundPage = true;
-				break;
-			}
 
-			// If the searched date is more recent than (after) the first (most recent) photo, ie.
-			// if searchedDate > date(first photo), then we have to get back a bit to more recent photos
-			else if (searchedDate > Long.valueOf(firstphoto.getAttribute("dateadded"))) {
-				if (page == 1) {
-					// If page == 1, then the searched date is later than last post on the group, so we actually found the date
-					logger.debug("There was no post later than the searched date");
-					foundPage = true;
+				// If the searched date is between the first (most recent) and the last (oldest) photos of the page, ie.
+				// older than the first photo and more recent than the last photo, ie.
+				// if date(last photo) <= searchedDate <= date(first photo), then we have found the right page!
+				else if (Long.valueOf(lastphoto.getAttribute("dateadded")) <= searchedDate && searchedDate <= Long.valueOf(firstphoto.getAttribute("dateadded"))) {
 					indexInPage = 0;
+					while (indexInPage < photosList.getLength()) {
+						Element photo = (Element) photosList.item((int) indexInPage);
+						if (searchedDate > Long.valueOf(photo.getAttribute("dateadded")))
+							break;
+						indexInPage++;
+						// When this inner loop is over, we have found the first photo, in the page, that is more recent than startDate
+					}
+					foundPage = true;
 					break;
 				}
-				if (page - increment > 0)
-					page -= increment;
-				logger.debug("searchedDate > " + firstphoto.getAttribute("dateadded") + ": going futur");
+
+				// If the searched date is more recent than (after) the first (most recent) photo, ie.
+				// if searchedDate > date(first photo), then we have to get back a bit to more recent photos
+				else if (searchedDate > Long.valueOf(firstphoto.getAttribute("dateadded"))) {
+					if (page == 1) {
+						// If page == 1, then the searched date is later than last post on the group, so we actually found the date
+						logger.debug("There was no post later than the searched date");
+						foundPage = true;
+						indexInPage = 0;
+						break;
+					}
+					if (page - increment > 0)
+						page -= increment;
+					logger.debug("searchedDate > " + firstphoto.getAttribute("dateadded") + ": going futur");
+				}
 			}
 
 			// Next turn in the dicotomy
@@ -1007,15 +1010,15 @@ private String getPhotoUrl(Element photoElement, ImageType type) {
 
 		String url = "http://farm" + farmId + ".static.flickr.com/" + serverId + "/" + id + "_" + secret;
 		switch (type) {
-		case SQUARE:
-			return url + "_s.jpg";
-		case THUMBNAIL:
-			return url + "_t.jpg";
-		case BIG:
-			return url + "_b.jpg";
-		case MEDIUM:
-		default:
-			return url + ".jpg";
+			case SQUARE:
+				return url + "_s.jpg";
+			case THUMBNAIL:
+				return url + "_t.jpg";
+			case BIG:
+				return url + "_b.jpg";
+			case MEDIUM:
+			default:
+				return url + ".jpg";
 		}
 	}
 

diff --git a/src/fm/flickr/stat/ProcessProbabilityPerWeekDayAndHour.java b/src/fm/flickr/stat/ProcessProbabilityPerWeekDayAndHour.java
@@ -108,6 +108,9 @@ private static void loadFileByDay(String date) throws ServiceException {
 	}
 
 	/**
+	 * 
+	 * Distribute the data loaded about upload time and number of uploads into 2 tables 
+	 * of the same data sumed by date and time slot
 	 * @param ps the output where to print results
 	 * @param dayOfWeek 1=Sunday to 7=Saturday 
 	 */
@@ -116,8 +119,6 @@ private static void computeStatistics(int dayOfWeek, PrintStream ps) {
 		Vector<Integer> postTimeDistrib = TimeStat.getPostTimeDistrib();
 		Vector<Long> uploadDistrib = DailyUploadsStat.getUploadDistribution();
 
-		// Print the results cut down by hour of day, from 0h to 23h
-		// Output format is: YYY-MM-DD HH:MM; nb of explored photos; nb of uploads; % of explored/posted  
 		for (int hour = 0; hour < 24; hour++) {
 
 			int curVal = exploredTable.get(dayOfWeek).get(hour);

diff --git a/src/fm/flickr/stat/README.txt b/src/fm/flickr/stat/README.txt
@@ -13,13 +13,21 @@ Read data between fm.flickr.stat.startdate and fm.flickr.stat.enddate, and compu
       "nb of explored photos in a group / total nb of photos uploaded to that group" during the same time slot.
       This takes a long time (lots of requests), to be activated cautiously...
 - upload: distribution of number of photos posted to Flickr in function of the hour of day
-- activity and anyphoto:
+- activity about explored photos:
+	- distribution of number of explored photos in function of the number of groups they are posted to
+	- distribution of number of explored photos in function of the number of times they were viewed
+	- distribution of number of explored photos in function of the number of comments they have
+	- distribution of number of explored photos in function of the number of favs they have
+	- distribution of number of explored photos in function of the owner's number of photos (data collected starting 15 Nov 2012)
+	- distribution of number of explored photos in function of the owner's number of contacts (data collected starting 15 Nov 2012)
+- activity about any unexplored photos: data collected starting 8 Jan. 2013 => for photos starting at 8 Dec. 2012:
 	- distribution of number of explored photos in function of the number of groups they are posted to
 	- distribution of number of explored photos in function of the number of times they were viewed
 	- distribution of number of explored photos in function of the number of comments they have
 	- distribution of number of explored photos in function of the number of favs they have
-	- distribution of number of explored photos in function of the owner's number of photos (starting 15/11/2012)
-	- distribution of number of explored photos in function of the owner's number of contacts (starting 15/11/2012)
+	- distribution of number of explored photos in function of the owner's number of photos
+	- distribution of number of explored photos in function of the owner's number of contacts
+
 
 ProcessMonthlyStats:
 ====================
@@ -39,8 +47,8 @@ by month, for categories group, tag, user, time, user, activity and upload:
 	- file monthly_distrib_view.csv: distribution of number of explored photos in function of the number of times they were viewed
 	- file monthly_distrib_comment.csv: distribution of number of explored photos in function of the number of comments they have
 	- file monthly_distrib_fav.csv: distribution of number of explored photos in function of the number of favs they have
-	- file monthly_distrib_owners_photo.csv shows the distribution of those users per number of photos that they have (starting 15/11/2012)
-	- file monthly_distrib_owners_contact.csv shows the distribution of those users per number of contacts that they have (starting 15/11/2012)
+	- file monthly_distrib_owners_photo.csv shows the distribution of users per number of photos that they have (starting 15/11/2012)
+	- file monthly_distrib_owners_contact.csv shows the distribution of users per number of contacts that they have (starting 15/11/2012)
 
 
 ProcessProbabilityPerWeekDayAndHour: