WIP snapshot

analyzers/__init__.py

@@ -51,7 +51,7 @@ __MAPPING__ = {
 		StoreRender
 	],
 	SimulationOrderAnalyzer: [
-		JSONRender,
+		#JSONRender,
 		# SimulationOrderRender,
 		SimulationGroupRender
 	]

analyzers/analyzer/__init__.py

@@ -50,7 +50,7 @@ class ResultStore:
 		:return:
 		"""
 		result = []
-		for key in self.store:
+		for key in sorted(self.store):
 			result += self.store[key]
 		return result
 

analyzers/analyzer/biogames.py

@@ -195,6 +195,7 @@ class ActivityMapper(Analyzer):
 			board_data = get_board_data(self.settings.source, self.instance_config_id, entry["sequence_id"],
 										entry["board_id"])
 			entry["extra_data"] = board_data
+			entry["extra_data"]["activity_type"] = self.last_board_type
 			entry['coordinate'] = self.new_coordinate()
 			self.timeline.append(entry)
 		return False
@@ -236,6 +237,8 @@ class ActivityMapper(Analyzer):
 		self.track['properties'].update(props)
 		self.tracks.append(self.track)
 		self.track = self.new_track(props['end_timestamp'])
+		if self.last_coordinate:
+			self.track['coordinates'].append(self.last_coordinate)
 
 	def new_track(self, timestamp):
 		return {"type": "LineString", "coordinates": [], "properties": {'start_timestamp': timestamp}}

analyzers/render/biogames.py

@@ -185,7 +185,13 @@ class SimulationOrderRender(Render):
 
 class SimulationGroupRender(Render):
 	def render(self, results: List[Result], name=None):
-		data = [r.get() for r in self.filter(results)]
+		#data = [r.get() for r in self.filter(results)]
+		data = []
+		for r in self.filter(results):
+			raw = r.get()
+			if len(raw) < 6:
+				raw = [0] + raw
+			data.append(raw)
 		print(name, len(data))
 		graph_plot(list(data), ylabel="simulation retries", title="sequential simulation retries", rotation=None, name=name)
 		#graph_fit(list(data), name=name)

biogames2.json

@@ -14,6 +14,7 @@
   "analyzers": {
     "analyzers": [
       "SimulationCategorizer",
+      "SimulationOrderAnalyzer",
       "ActivityMapper"
     ]
   },

log_analyzer.py

@@ -63,17 +63,18 @@ if __name__ == '__main__':
 	# "fec57041458e6cef98652df625", ]
 	log_ids = []
 	# with open("/home/clemens/git/ma/test/filtered") as src:
-	with open("/home/clemens/git/ma/test/filtered") as src:
-		for line in src:
-			line = line.strip()
-			log_ids.append(line)
-	store: ResultStore = ResultStore()
-	for log_id in log_ids:
-		for analysis in process_log(log_id, settings):
-			log.info("* Result for " + analysis.name())
-			# print(analysis.result())
-			# print(analysis.render())
-			analysis.result(store)
+	if False:
+		with open("/home/clemens/git/ma/test/filtered_5_actions") as src:
+			for line in src:
+				line = line.strip()
+				log_ids.append(line)
+		store: ResultStore = ResultStore()
+		for log_id in log_ids:
+			for analysis in process_log(log_id, settings):
+				log.info("* Result for " + analysis.name())
+				# print(analysis.result())
+				# print(analysis.render())
+				analysis.result(store)
 	if False:
 		for r in get_renderer(analyzers.LocomotionActionAnalyzer):
 			r().render(store.get_all())
@@ -120,8 +121,10 @@ if __name__ == '__main__':
 					writer.writerow(line)
 
 	if True:
-		#json.dump(store.serializable(), open("new.json", "w"), indent=1)
+		# json.dump(store.serializable(), open("new.json", "w"), indent=1)
 		from collections import defaultdict
+		import matplotlib.pyplot as plt
+		from util.meta_temp import CONFIG_NAMES
 
 		keys = [
 			"simu",
@@ -130,81 +133,294 @@ if __name__ == '__main__':
 			"audio",
 			"video",
 			"other",
-			"map"
+			"map",
+			# "error"
 		]
-		import matplotlib.pyplot as plt
-		#results = []
 
-		places = defaultdict(list)
 
-		for log in store.get_all():
-			result = defaultdict(lambda: 0)
-			for i in log.get()['track']:
-				duration = i['properties']['end_timestamp'] - i['properties']['start_timestamp']
-				result[i['properties']['activity_type']] += duration
-			print(json.dumps(result, indent=4))
-			total = sum(result.values())
-			print(total)
-			percentage = defaultdict(lambda :0)
-			minutes = defaultdict(lambda:0)
-			for i in result:
-				percentage[i]= result[i]/total
-				minutes[i] = result[i]/60_000
-			print(json.dumps(percentage,indent=4))
-			if not 'error' in result:
-				#places[log.get()['instance']].append(percentage)
-				places[log.get()['instance']].append(minutes)
+		def get_data(store, relative_values=True, sort=True, show_errors=False):
+			places = defaultdict(list)
 
-		for place in places:
-			places[place] = sorted(places[place], key=lambda item:item['map'])
+			for log in store.get_all():
+				if not log.analysis() == analyzers.ActivityMapper:
+					continue
+				result = defaultdict(lambda: 0)
+				for i in log.get()['track']:
+					duration = i['properties']['end_timestamp'] - i['properties']['start_timestamp']
+					result[i['properties']['activity_type']] += duration
+				print(json.dumps(result, indent=4))
+				total = sum(result.values())
+				print(total)
+				percentage = defaultdict(lambda: 0)
+				minutes = defaultdict(lambda: 0)
+				for i in result:
+					percentage[i] = result[i] / total
+					minutes[i] = result[i] / 60_000
+				print(json.dumps(percentage, indent=4))
+				if not 'error' in result or show_errors:
+					if relative_values:
+						places[log.get()['instance']].append(percentage)
+					else:
+						places[log.get()['instance']].append(minutes)
+			if sort:
+				for place in places:
+					places[place] = sorted(places[place], key=lambda item: item['map'])
+			return places
 
-		dummy = [0]*len(keys)
-		results = []
-		sites = []
-		from util.meta_temp import CONFIG_NAMES
-		for i in places:
-			for j in places[i]:
-				ordered = []
+
+		from shapely.geometry import LineString
+		from shapely.ops import transform
+		from functools import partial
+		import pyproj
+
+
+		def calc_distance(coordinates):
+			track = LineString(coordinates)
+			project = partial(
+				pyproj.transform,
+				pyproj.Proj(init='EPSG:4326'),
+				pyproj.Proj(init='EPSG:32633'))
+			return transform(project, track).length
+
+		whitelist = ['16fc3117-61db-4f50-b84f-81de6310206f', '5e64ce07-1c16-4d50-ac4e-b3117847ea43',
+					 '90278021-4c57-464e-90b1-d603799d07eb', 'ff8f1e8f-6cf5-4a7b-835b-5e2226c1e771']
+
+
+		def get_data_distance(store, relative_values=True, sort=True, show_errors=False):
+			places = defaultdict(list)
+
+			for log in store.get_all():
+				if not log.analysis() == analyzers.ActivityMapper:
+					continue
+				result = defaultdict(lambda: 0)
+				for i in log.get()['track']:
+					coords = i['coordinates']
+					if len(coords) > 1:
+						distance = calc_distance(coords)
+						result[i['properties']['activity_type']] += distance
+				total = sum(result.values())
+				percentage = defaultdict(lambda: 0)
+				for i in result:
+					if not total == 0:
+						percentage[i] = result[i] / total
+				if not 'error' in result or show_errors:
+					if relative_values:
+						places[log.get()['instance']].append(percentage)
+					else:
+						places[log.get()['instance']].append(result)
+			if sort:
+				for place in places:
+					places[place] = sorted(places[place], key=lambda item: item['map'])
+			return places
+
+
+		def get_all_data(store, sort=False, relative=True):
+			places = defaultdict(list)
+			simu_distribution = defaultdict(lambda: 0)
+			#divisiors = {"time":60_000, "space":1000000}
+			for log in store.get_all():
+				if not log.analysis() == analyzers.ActivityMapper:
+					continue
+				result = defaultdict(lambda: defaultdict(lambda: 0))
+				for i in log.get()['track']:
+					coords = i['coordinates']
+					if len(coords) > 1:
+						distance = calc_distance(coords)
+					else:
+						distance = 0.0
+					result["space"][i['properties']['activity_type']] += distance
+					duration = i['properties']['end_timestamp'] - i['properties']['start_timestamp']
+					result["time"][i['properties']['activity_type']] += duration
+				total_space = sum(result["space"].values())
+				total_time = sum(result["time"].values())
+				percentage = defaultdict(lambda: defaultdict(lambda: 0))
+				total = defaultdict(lambda: defaultdict(lambda: 0))
+				for i in result["space"]:
+					if not total_space == 0:
+						percentage[i]["space"] = result["space"][i] / total_space
+					else:
+						percentage[i]["space"] = 0
+					if not total_time == 0:
+						percentage[i]["time"] = result["time"][i] / total_time
+					else:
+						percentage[i]["time"] = 0
+					for t in ("space","time"):
+						#total[i][t] += (result[t][i] / divisiors[t])
+						total[i][t] += result[t][i]
+				print(percentage)
+				if not 'error' in result:
+					if relative:
+						value = percentage
+					else:
+						value = total
+					places[log.get()['instance']].append(value)
+					simus = defaultdict(lambda :0)
+					for item in log.get()['boards']:
+						if item["extra_data"]["activity_type"]=="simu":
+							simus[item["board_id"]] += 1
+					simu_distribution[len(simus)]+=1
+
+			if sort:
+				for place in places:
+					places[place] = sorted(places[place], key=lambda item: item['map']['time'])
+			print(simu_distribution)
+			return places
+
+
+		def stack_data(keys, places, type="time"):
+			divisiors = {"time": 60_000, "space": 1000}
+			divisiors = {"time": 1, "space": 1}
+			dummy = [0] * len(keys)
+			results = []
+			sites = []
+			for i in sorted(places):
+				if not i in whitelist:
+					continue
+				for j in places[i]:
+					ordered = []
+					for k in keys:
+						if k in j:
+							ordered.append(j[k][type]/divisiors[type])
+						else:
+							ordered.append(0)
+					print(sum(ordered))
+					if sum(ordered) > 0.9:
+						#print(sum(ordered), 1-sum(ordered))
+						#if sum(ordered)<1:
+						#	ordered[-2] = 1-sum(ordered[:-2], ordered[-1])
+						results.append(ordered)
+				results.append(dummy)
+				sites.append(CONFIG_NAMES[i] if i in CONFIG_NAMES else "---")
+			return results, sites
+
+
+		def plot_data(places, keys):
+			results, sites = stack_data(keys, places)
+			dpi=86.1
+			plt.figure(figsize=(1280/dpi, 720/dpi))
+			size = len(results)
+			print("{} elements total".format(size))
+			ind = np.arange(size)
+			width = 1
+			# print(results)
+			data = list(zip(*results))
+			# print(data)
+			lines = []
+			bottom = [0] * size
+			plt.ticklabel_format(useMathText=False)
+			for i in range(0, len(data)):
+				lines.append(plt.bar(ind, data[i], bottom=bottom, width=width)[0])
+				for k, x in enumerate(data[i]):
+					bottom[k] += x
+			plt.legend(lines, keys)
+			plt.title(", ".join(sites))
+			#plt.show()
+			dpi=86
+			plt.savefig("time_rel_{}.png".format(size), dpi=dpi,bbox_inches="tight")
+
+
+		colors = {
+			"simu": "blue",
+			"question": "orange",
+			"image": "green",
+			"audio": "red",
+			"video": "purple",
+			"other": "brown",
+			"map": "violet",
+			# "error":"grey"
+		}
+		markers = [".", "o", "x", "s", "*", "D", "p", ",", "<", ">", "^", "v", "1", "2", "3", "4"]
+
+
+		def plot_time_space(time_data, space_data, keys):
+			# assuming time_data and space_data are in same order!
+			marker = 0
+			for id in time_data:
 				for k in keys:
-					ordered.append(j[k])
-				results.append(ordered)
-			results.append(dummy)
-			sites.append(CONFIG_NAMES[i] if i in CONFIG_NAMES else "---")
+					for i in range(len(time_data[id])):
+						print(time_data[id][i][k], space_data[id][i][k])
+						plt.plot(time_data[id][i][k], space_data[id][i][k], color=colors[k], marker=markers[marker])
+				marker += 1
+			plt.show()
 
 
-		size = len(results)
-		ind = np.arange(size)
-		width=0.9
-		print(results)
-		data = list(zip(*results))
-		print(data)
-		lines = []
-		bottom = [0]*len(results)
-		for i in range(0, len(data)):
-			lines.append(plt.bar(ind,data[i], bottom=bottom, width=width)[0])
-			for k,x in enumerate(data[i]):
-				bottom[k] += x
-		plt.legend(lines, keys)
-		plt.title(", ".join(sites))
-		plt.show()
+		# plt.cla()
+		# plt.clf()
+		# plt.close()
 
-	#size = len(results)
-	#ind = np.arange(size)
-	#width = 0.9
-	#print(results)
-	#data = list(zip(*results))
-	#print(data)
-	#lines = []
-	#bottom = [0] * len(results)
-	#for i in range(0, len(data)):
-	#	lines.append(plt.bar(ind, data[i], bottom=bottom, width=width)[0])
-	#	for k, x in enumerate(data[i]):
-	#		bottom[k] += x
-	#plt.legend(lines, keys)
-	#plt.title("Zwei Spiele in Filderstadt (t1=237min; t2=67min)")
-	#plt.show()
+		def plot_time_space_rel(combined, keys):
+			groups = defaultdict(list)
+			keys = list(keys)
+			keys.remove("other")
+			ids = []
+			for k in keys:
+				for id in sorted(combined):
+					if id not in whitelist:
+						continue
+					if not id in ids:
+						ids.append(id)
+					group = 0.0
+					count = 0
+					for item in combined[id]:
+						if k in item:
+							time = item[k]["time"]/1000
+							distance = item[k]["space"]
+							if time > 0:
+								group += (distance / time)
+								count+=1
+							else:
+								print("div by zero", distance, time)
+					if count > 0:
+						groups[k].append(group/count)
+					else:
+						groups[k].append(0.0)
+			print(ids)
+			ind = np.arange(len(ids))
+			width = .7 / len(groups)
+			print(ind)
+			print(json.dumps(groups, indent=1))
+			bars = []
+			dpi=10
+			plt.figure(figsize=(1280/dpi, 720/dpi))
+			fig, ax = plt.subplots()
+			for k in groups:
+				print(groups[k])
+				if not len(groups[k]):
+					groups[k].append(0)
+				ind = ind + (width)
+				bars.append(ax.bar((ind + width*len(groups)/2), groups[k], width, color=colors[k]))
+			ax.set_xticks(ind + width / 2)
+			ax.set_xticklabels(list([CONFIG_NAMES[i] if i in CONFIG_NAMES else "---" for i in ids]))
+			plt.legend(bars, keys)
+			print(combined.keys(), ids)
+			print([CONFIG_NAMES[i] if i in CONFIG_NAMES else "---" for i in ids])
+			#plt.show()
+			dpi=100
+			plt.savefig("speed.png", dpi=dpi)
 
 
+		# spatial_data = get_data_distance(store,relative_values=False)
+		# temporal_data = get_data(store,relative_values=False)
+		# spatial_data_rel = get_data_distance(store,relative_values=True)
+		# temporal_data_rel = get_data(store,relative_values=True)
+		#temporal_data_rel = json.load(open("temporal_rel.json"))
+		#spatial_data_rel = json.load(open("spatial_rel.json"))
+		# import IPython
+		# IPython.embed()
+
+		#print(json.dumps(get_all_data(store)))
+		#json.dump(get_all_data(store), open("combined.json", "w"))
+		#combined = get_all_data(store, sort=True, relative=True)
+		#json.dump(combined, open("combined_rel.json", "w"))
+		#combined = json.load(open("combined_rel.json"))
+		combined = json.load(open("combined_total.json"))
+		plot_time_space_rel(combined, keys)
+
+		#plot_time_space_rel(temporal_data_rel, spatial_data_rel, keys)
+
+		#plot_data(combined, keys)
+	# plot_data(get_data_distance(store,relative_values=False), keys)
+
 
 # for analyzers in analyzers:
 #	if analyzers.name() in ["LogEntryCount", "ActionSequenceAnalyzer"]:

requirements.txt

@@ -4,4 +4,5 @@ matplotlib==2.1.0
 osmnx==0.6
 networkx==2.0
 pydot==1.2.3
-scipy==1.0.0
+scipy==1.0.0
+ipython==6.2.1