From cbe414232cc2168f10f84ce3d2bcd9b752605360 Mon Sep 17 00:00:00 2001
From: Clemens Klug <clemens.klug@uni-bamberg.de>
Date: Tue, 15 Oct 2019 14:42:51 +0200
Subject: [PATCH] improve time tracking of tasks and CSV formatting

---
 analysis/analyzers/analyzer/biogames.py | 18 ++++++++++++------
 analysis/util/output.py                 | 19 ++++++++++++++++---
 2 files changed, 28 insertions(+), 9 deletions(-)

diff --git a/analysis/analyzers/analyzer/biogames.py b/analysis/analyzers/analyzer/biogames.py
index 970cb62..3dc0d9e 100644
--- a/analysis/analyzers/analyzer/biogames.py
+++ b/analysis/analyzers/analyzer/biogames.py
@@ -365,6 +365,7 @@ class BiogamesTasks(Analyzer):
 		super().__init__(settings)
 		self.settings: LogSettings = settings
 		self.tasks = {}
+		self.first_board_view = {}
 		self.last_board = None
 		self.instance_config_id: str = None
 
@@ -372,11 +373,15 @@ class BiogamesTasks(Analyzer):
 		if self.instance_config_id is None:
 			if entry[self.settings.type_field] in self.settings.custom['instance_start']:
 				self.instance_config_id = json_path(entry, self.settings.custom['instance_config_id'])
-		if self.is_task(entry) and self.last_board:
-			entry['__duration'] = entry['timestamp'] - self.last_board['timestamp']
-			self.tasks[self.ids()] = entry
+		if self.is_task(entry):
+			task_id = entry['answers']['@id']
+			if task_id not in self.first_board_view:
+				logger.error("task_id is not in self.first_board_view!", task_id, entry)
+			else:
+				entry['__duration'] = entry['timestamp'] - self.first_board_view[task_id]['timestamp']
+				self.tasks[self.ids(task_id)] = entry
 		if self.is_board(entry):
-			self.last_board = entry
+			self.first_board_view[entry['board_id']] = entry
 		return False
 
 	def result(self, store: ResultStore, name=None) -> None:
@@ -388,8 +393,9 @@ class BiogamesTasks(Analyzer):
 					results[ids] = {"duration": task['__duration'], "result": action['increment']}
 		store.add(Result(type(self), results))
 
-	def ids(self):
-		return f"{self.instance_config_id}_{self.last_board['sequence_id']}_{self.last_board['board_id']}"
+	def ids(self, task_id):
+		task = self.first_board_view[task_id]
+		return f"{self.instance_config_id}_{task['sequence_id']}_{task['board_id']}"
 
 	def is_task(self, entry) -> bool:
 		return entry['@class'] in self.DATA_CLASSES
diff --git a/analysis/util/output.py b/analysis/util/output.py
index 1ad705f..1908773 100644
--- a/analysis/util/output.py
+++ b/analysis/util/output.py
@@ -1,14 +1,27 @@
 from datetime import datetime as dt
 
+SEP = "\",\""
+LS = "\""
+LE = "\""
+NL = LS + "\n" + LE
+
+
+
 def flat_dict_to_csv(data):
 	keys = set()
 	for i in data:
 		keys = keys.union(set(i.keys()))
 	keys = sorted(keys)
-	out = ",".join(keys)
+	out = SEP.join(keys)
 	for i in data:
-		out += "\n" + ",".join([str(i.get(j, "")) for j in keys])
-	return out
+		out += NL + SEP.join([escape(i.get(j, "")) for j in keys])
+	return LS + out + LE
+
+
+def escape(value):
+	val = str(value)
+	val = val.replace(".", ",")
+	return val
 
 def pretty_ts(timestamp, fmt="%Y-%m-%d %H:%M:%S"):
 	d = dt.fromtimestamp(int(timestamp)/1000.0)