def getReviewScoreInfo(inputFile):
"""
review_score.csv
data format:
review ID | field ID | score
File has header
e.g. 1,1,3 - score (can be negative)
1,2,5 - confidence
1,3,no - recommended
"""
parsedResult = {}
parsedCSV = parseCSVFile(inputFile)
headers = parsedCSV[:1][0]
print(headers)
schema = ['review #', 'field #', 'score']
schemaIndex = [headers.index(header) for header in schema]
lines = parsedCSV[1:]
lines = [ele for ele in lines if ele]
scores = []
confidences = []
isRecommended = []
scores = [int(line[2]) for line in lines if int(line[1]) == 1]
confidences = [int(line[2]) for line in lines if int(line[1]) == 2]
isRecommended = [str(line[2]).replace("\r", "") for line in lines if int(line[1]) == 3]
parsedResult['yesPercentage'] = float(isRecommended.count('yes')) / len(isRecommended)
parsedResult['meanScore'] = sum(scores) / float(len(scores))
parsedResult['meanConfidence'] = sum(confidences) / float(len(confidences))
parsedResult['totalReview'] = len(confidences)
return {'infoType': 'reviewScore', 'infoData': parsedResult}
def getAuthorInfo(inputFile):
"""
author.csv: header row, author names with affiliations, countries, emails
data format:
submission ID | f name | s name | email | country | affiliation | page | person ID | corresponding?
"""
parsedResult = {}
parsedCSV = parseCSVFile(inputFile)
headers = parsedCSV[:1][0]
schema = ['submission #', 'first name', 'last name', 'email', 'country', 'organization', 'Web page', 'person #', 'corresponding?']
schemaIndex = [headers.index(header) for header in schema]
lines = parsedCSV[1:]
lines = [ele for ele in lines if ele]
authorList = []
authorModels = []
for authorInfo in lines:
authorList.append({'name': authorInfo[schemaIndex[1]] + " " + authorInfo[schemaIndex[2]], 'country': authorInfo[schemaIndex[4]], 'affiliation': authorInfo[schemaIndex[5]]})
authorModels.append(Author(submission = Submission.objects.filter(submission_no= int(authorInfo[schemaIndex[0]])).first(), name = authorInfo[schemaIndex[1]] + " " + authorInfo[schemaIndex[2]], country = authorInfo[schemaIndex[4]], organisation = authorInfo[schemaIndex[5]]))
Author.objects.all().delete()
Author.objects.bulk_create(authorModels)
authors = [ele['name'] for ele in authorList if ele] # adding in the if ele in case of empty strings; same applies below
topAuthors = Counter(authors).most_common(10)
parsedResult['topAuthors'] = {'labels': [ele[0] for ele in topAuthors], 'data': [ele[1] for ele in topAuthors]}
countries = [ele['country'] for ele in authorList if ele]
topCountries = Counter(countries).most_common(10)
parsedResult['topCountries'] = {'labels': [ele[0] for ele in topCountries], 'data': [ele[1] for ele in topCountries]}
affiliations = [ele['affiliation'] for ele in authorList if ele]
topAffiliations = Counter(affiliations).most_common(10)
parsedResult['topAffiliations'] = {'labels': [ele[0] for ele in topAffiliations], 'data': [ele[1] for ele in topAffiliations]}
topAuthorInCountry ={}
for country in countries:
myAuthorInCountry = [ele['name'] for ele in authorList if ele['country'] == country]
myAuthorInCountry = Counter(myAuthorInCountry).most_common(10)
topAuthorInCountry[country] = {'labels': [ele[0] for ele in myAuthorInCountry], 'data': [ele[1] for ele in myAuthorInCountry]}
parsedResult['topAuthorInCountry'] = topAuthorInCountry
return {'infoType': 'author', 'infoData': parsedResult}
def getReviewInfo(inputFile):
"""
review.csv
data format:
review ID | paper ID? | reviewer ID | reviewer name | unknown | text | scores | overall score | unknown | unknown | unknown | unknown | date | time | recommend?
File has NO header
score calculation principles:
Weighted Average of the scores, using reviewer's confidence as the weights
recommended principles:
Yes: 1; No: 0; weighted average of the 1 and 0's, also using reviewer's confidence as the weights
"""
parsedResult = {}
parsedCSV = parseCSVFile(inputFile)
headers = parsedCSV[:1][0]
print(headers)
schema = ['review ID', 'paper ID', 'reviewer ID', 'reviewer name', 'unknown', 'text', 'scores', 'overall score', 'unknown', 'unknown', 'unknown', 'unknown', 'date', 'time','recommend']
schemaIndex = [headers.index(header) for header in schema]
lines = parsedCSV[1:]
lines = [ele for ele in lines if ele]
evaluation = [str(line[schemaIndex[6]]).replace("\r", "") for line in lines]
submissionIDs = set([str(line[schemaIndex[1]]) for line in lines])
reviewerName = [line[schemaIndex[3]] for line in lines ]
topReviewer = Counter(reviewerName).most_common(10)
scoreList = []
recommendList = []
confidenceList = []
submissionIDReviewMap = {}
# Idea: from -3 to 3 (min to max scores possible), every 0.25 will be a gap
scoreDistributionCounts = [0] * int((3 + 3) / 0.25)
recommendDistributionCounts = [0] * int((1 - 0) / 0.1)
scoreDistributionLabels = [" ~ "] * len(scoreDistributionCounts)
recommendDistributionLabels = [" ~ "] * len(recommendDistributionCounts)
for index, col in enumerate(scoreDistributionCounts):
scoreDistributionLabels[index] = str(-3 + 0.25 * index) + " ~ " + str(-3 + 0.25 * index + 0.25)
for index, col in enumerate(recommendDistributionCounts):
recommendDistributionLabels[index] = str(0 + 0.1 * index) + " ~ " + str(0 + 0.1 * index + 0.1)
for submissionID in submissionIDs:
reviews = [str(line[schemaIndex[6]]).replace("\r", "") for line in lines if str(line[schemaIndex[1]]) == submissionID]
# print reviews
confidences = [float(review.split("\n")[1].split(": ")[1]) for review in reviews]
scores = [float(review.split("\n")[0].split(": ")[1]) for review in reviews]
confidenceList.append(sum(confidences) / len(confidences))
# recommends = [1.0 for review in reviews if review.split("\n")[2].split(": ")[1] == "yes" else 0.0]
try:
recommends = map(lambda review: 1.0 if review.split("\n")[2].split(": ")[1] == "yes" else 0.0, reviews)
except:
recommends = [0.0 for n in range(len(reviews))]
weightedScore = sum(x * y for x, y in zip(scores, confidences)) / sum(confidences)
weightedRecommend = sum(x * y for x, y in zip(recommends, confidences)) / sum(confidences)
scoreColumn = min(int((weightedScore + 3) / 0.25), 23)
recommendColumn = min(int((weightedRecommend) / 0.1), 9)
scoreDistributionCounts[scoreColumn] += 1
recommendDistributionCounts[recommendColumn] += 1
submissionIDReviewMap[submissionID] = {'score': weightedScore, 'recommend': weightedRecommend}
scoreList.append(weightedScore)
recommendList.append(weightedRecommend)
topExperts = {}
topTen = []
topTenCount =[]
topTenAverage=[]
for line in lines:
if line[3] not in topExperts:
topExperts[str(line[schemaIndex[3]])] = [1, int(line[schemaIndex[4]])]
else:
topExperts[str(line[schemaIndex[3]])] = [topExperts[str(line[schemaIndex[3]])][0]+1,topExperts[str(line[schemaIndex[3]])][1]+int(line[schemaIndex[4]])]
topExpertsArr = []
for name, value in topExperts.items():
topExpertsArr.append([name, value[0], round(value[1] * 1.0 / value[0] * 1.0, 1)])
def takeThird(elem):
return elem[2]
topExpertsArr.sort(key=takeThird, reverse = True)
topTen = [ele[0] for ele in topExpertsArr[:10]]
topTenCount = [ele[1] for ele in topExpertsArr[:10]]
topTenAverage = [ele[2] for ele in topExpertsArr[:10]]
parsedResult['IDReviewMap'] = submissionIDReviewMap
parsedResult['scoreList'] = scoreList
parsedResult['meanScore'] = sum(scoreList) / len(scoreList)
parsedResult['meanRecommend'] = sum(recommendList) / len(recommendList)
parsedResult['meanConfidence'] = sum(confidenceList) / len(confidenceList)
parsedResult['recommendList'] = recommendList
parsedResult['scoreDistribution'] = {'labels': scoreDistributionLabels, 'counts': scoreDistributionCounts}
parsedResult['recommendDistribution'] = {'labels': recommendDistributionLabels, 'counts': recommendDistributionCounts}
parsedResult['topReviewers'] = {'labels': [ele[0] for ele in topReviewer], 'data': [ele[1] for ele in topReviewer]}
parsedResult['topExperts'] = {'labels': topTen, 'counts': topTenCount, 'averages': topTenAverage}
return {'infoType': 'review', 'infoData': parsedResult}
def getSubmissionInfo(inputFile):
"""
submission.csv
data format:
submission ID | track ID | track name | title | authors | submit time | last update time | form fields | keywords | decision | notified | reviews sent | abstract
File has header
"""
parsedResult = {}
parsedCSV = parseCSVFile(inputFile)
headers = parsedCSV[:1][0]
print(headers)
schema = ['#', 'track #', 'track name', 'title', 'authors', 'submitted', 'last updated', 'form fields', 'keywords', 'decision', 'notified', 'reviews sent', 'abstract']
schemaIndex = [headers.index(header) for header in schema]
lines = parsedCSV[1:]
lines = [ele for ele in lines if ele]
submissionModels = []
for submissionInfo in lines:
submissionModels.append(Submission(submission_no = int(submissionInfo[schemaIndex[0]]), track_no = int(submissionInfo[schemaIndex[1]]),decision = submissionInfo[schemaIndex[9]]))
Submission.objects.all().delete()
Submission.objects.bulk_create(submissionModels)
acceptedAffiliations = [a.organisation for a in Author.objects.filter(submission__decision = 'accept')]
topAcceptedAffiliations = Counter(acceptedAffiliations).most_common(10)
topAcceptedAffiliations = {'names': [ele[0] for ele in topAcceptedAffiliations], 'counts': [ele[1] for ele in topAcceptedAffiliations]}
acceptedSubmission = [line for line in lines if str(line[schemaIndex[9]]) == 'accept']
rejectedSubmission = [line for line in lines if str(line[schemaIndex[9]]) == 'reject']
trackName = [line[schemaIndex[2]] for line in lines ]
acceptanceRate = float(len(acceptedSubmission)) / len(lines)
submissionTimes = [parseSubmissionTime(str(ele[schemaIndex[5]])) for ele in lines]
lastEditTimes = [parseSubmissionTime(str(ele[schemaIndex[6]])) for ele in lines]
submissionTimes = Counter(submissionTimes)
lastEditTimes = Counter(lastEditTimes)
timeStamps = sorted([k for k in submissionTimes])
lastEditStamps = sorted([k for k in lastEditTimes])
submittedNumber = [0 for n in range(len(timeStamps))]
lastEditNumber = [0 for n in range(len(lastEditStamps))]
timeSeries = []
lastEditSeries = []
for index, timeStamp in enumerate(timeStamps):
if index == 0:
submittedNumber[index] = submissionTimes[timeStamp]
else:
submittedNumber[index] = submissionTimes[timeStamp] + submittedNumber[index - 1]
timeSeries.append({'x': timeStamp, 'y': submittedNumber[index]})
for index, lastEditStamp in enumerate(lastEditStamps):
if index == 0:
lastEditNumber[index] = lastEditTimes[lastEditStamp]
else:
lastEditNumber[index] = lastEditTimes[lastEditStamp] + lastEditNumber[index - 1]
lastEditSeries.append({'x': lastEditStamp, 'y': lastEditNumber[index]})
# timeSeries = {'time': timeStamps, 'number': submittedNumber}
# lastEditSeries = {'time': lastEditStamps, 'number': lastEditNumber}
acceptedKeywords = [str(ele[8]).lower().replace("\r", "").split("\n") for ele in acceptedSubmission]
acceptedKeywords = [ele for item in acceptedKeywords for ele in item]
acceptedKeywordMap = {k : v for k, v in Counter(acceptedKeywords).iteritems()}
# acceptedKeywordList = [[ele[0], ele[1]] for ele in Counter(acceptedKeywords).most_common(20)]
acceptedKeywordList = {'names': [ele[0] for ele in Counter(acceptedKeywords).most_common(20)], 'counts': [ele[1] for ele in Counter(acceptedKeywords).most_common(20)]}
rejectedKeywords = [str(ele[8]).lower().replace("\r", "").split("\n") for ele in rejectedSubmission]
rejectedKeywords = [ele for item in rejectedKeywords for ele in item]
rejectedKeywordMap = {k : v for k, v in Counter(rejectedKeywords).iteritems()}
rejectedKeywordList = [[ele[0], ele[1]] for ele in Counter(rejectedKeywords).most_common(20)]
allKeywords = [str(ele[schemaIndex[8]]).lower().replace("\r", "").split("\n") for ele in lines]
allKeywords = [ele for item in allKeywords for ele in item]
allKeywordMap = {'names': [ele[0] for ele in Counter(allKeywords).most_common(20)], 'counts': [ele[1] for ele in Counter(allKeywords).most_common(20)]}
# allKeywordList = [[ele[0], ele[1]] for ele in Counter(allKeywords).most_common(20)]
allKeywordList = {'names': [ele[0] for ele in Counter(allKeywords).most_common(20)], 'counts': [ele[1] for ele in Counter(allKeywords).most_common(20)]}
tracks = set([str(ele[schemaIndex[2]]) for ele in lines])
paperGroupsByTrack = {track : [line for line in lines if str(line[schemaIndex[2]]) == track] for track in tracks}
keywordsGroupByTrack = {}
acceptanceRateByTrack = {}
comparableAcceptanceRate = {}
topAuthorsByTrack = {}
# Obtained from the JCDL.org website: past conferences
comparableAcceptanceRate['year'] = [2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018]
comparableAcceptanceRate['Full Papers'] = [0.29, 0.28, 0.27, 0.29, 0.29, 0.30, 0.29, 0.30]
comparableAcceptanceRate['Short Papers'] = [0.29, 0.37, 0.31, 0.31, 0.32, 0.50, 0.35, 0.32]
for track, papers in paperGroupsByTrack.iteritems():
keywords = [str(ele[8]).lower().replace("\r", "").split("\n") for ele in papers]
keywords = [ele for item in keywords for ele in item]
# keywordMap = {k : v for k, v in Counter(keywords).iteritems()}
keywordMap = [[ele[0], ele[1]] for ele in Counter(keywords).most_common(20)]
keywordsGroupByTrack[track] = {'names': [ele[0] for ele in Counter(keywords).most_common(20)], 'counts': [ele[1] for ele in Counter(keywords).most_common(20)]}
acceptedPapersPerTrack = [ele for ele in papers if str(ele[9]) == 'accept']
acceptanceRateByTrack[track] = float(len(acceptedPapersPerTrack)) / len(papers)
acceptedPapersThisTrack = [paper for paper in papers if str(paper[9]) == 'accept']
acceptedAuthorsThisTrack = [str(ele[4]).replace(" and ", ", ").split(", ") for ele in acceptedPapersThisTrack]
acceptedAuthorsThisTrack = [ele for item in acceptedAuthorsThisTrack for ele in item]
topAcceptedAuthorsThisTrack = Counter(acceptedAuthorsThisTrack).most_common(10)
topAuthorsByTrack[track] = {'names': [ele[0] for ele in topAcceptedAuthorsThisTrack], 'counts': [ele[1] for ele in topAcceptedAuthorsThisTrack]}
if track == "Full Papers" or track == "Short Papers":
comparableAcceptanceRate[track].append(float(len(acceptedPapersPerTrack)) / len(papers))
acceptedAuthors = [str(ele[4]).replace(" and ", ", ").split(", ") for ele in acceptedSubmission]
acceptedAuthors = [ele for item in acceptedAuthors for ele in item]
topAcceptedAuthors = Counter(acceptedAuthors).most_common(10)
topAcceptedAuthorsMap = {'names': [ele[0] for ele in topAcceptedAuthors], 'counts': [ele[1] for ele in topAcceptedAuthors]}
# topAcceptedAuthors = {ele[0] : ele[1] for ele in Counter(acceptedAuthors).most_common(10)}
topTrack = Counter(trackName).most_common(8)
parsedResult['acceptanceRate'] = acceptanceRate
parsedResult['overallKeywordMap'] = allKeywordMap
parsedResult['overallKeywordList'] = allKeywordList
parsedResult['acceptedKeywordMap'] = acceptedKeywordMap
parsedResult['acceptedKeywordList'] = acceptedKeywordList
parsedResult['rejectedKeywordMap'] = rejectedKeywordMap
parsedResult['rejectedKeywordList'] = rejectedKeywordList
parsedResult['keywordsByTrack'] = keywordsGroupByTrack
parsedResult['acceptanceRateByTrack'] = acceptanceRateByTrack
parsedResult['topAcceptedAuthors'] = topAcceptedAuthorsMap
parsedResult['topAuthorsByTrack'] = topAuthorsByTrack
parsedResult['timeSeries'] = timeSeries
parsedResult['lastEditSeries'] = lastEditSeries
parsedResult['comparableAcceptanceRate'] = comparableAcceptanceRate
parsedResult['topTracks'] = {'labels': [ele[0] for ele in topTrack], 'data': [ele[1] for ele in topTrack]}
parsedResult['topAcceptedAffiliations'] = topAcceptedAffiliations
# parsedResult['myAllKeywords'] = myAllKeywordsMap
return {'infoType': 'submission', 'infoData': parsedResult}
acceptedAuthors = [str(ele[4]).replace(" and ", ", ").split(", ") for ele in acceptedSubmission]
acceptedAuthors = [ele for item in acceptedAuthors for ele in item]
topAcceptedAuthors = Counter(acceptedAuthors).most_common(10)
topAcceptedAuthorsMap = {'names': [ele[0] for ele in topAcceptedAuthors], 'counts': [ele[1] for ele in topAcceptedAuthors]}
# topAcceptedAuthors = {ele[0] : ele[1] for ele in Counter(acceptedAuthors).most_common(10)}
topTrack = Counter(trackName).most_common(8)
parsedResult['acceptanceRate'] = acceptanceRate
parsedResult['overallKeywordMap'] = allKeywordMap
parsedResult['overallKeywordList'] = allKeywordList
parsedResult['acceptedKeywordMap'] = acceptedKeywordMap
parsedResult['acceptedKeywordList'] = acceptedKeywordList
parsedResult['rejectedKeywordMap'] = rejectedKeywordMap
parsedResult['rejectedKeywordList'] = rejectedKeywordList
parsedResult['keywordsByTrack'] = keywordsGroupByTrack
parsedResult['acceptanceRateByTrack'] = acceptanceRateByTrack
parsedResult['topAcceptedAuthors'] = topAcceptedAuthorsMap
parsedResult['topAuthorsByTrack'] = topAuthorsByTrack
parsedResult['timeSeries'] = timeSeries
parsedResult['lastEditSeries'] = lastEditSeries
parsedResult['comparableAcceptanceRate'] = comparableAcceptanceRate
parsedResult['topTracks'] = {'labels': [ele[0] for ele in topTrack], 'data': [ele[1] for ele in topTrack]}
parsedResult['topAcceptedAffiliations'] = topAcceptedAffiliations
# parsedResult['myAllKeywords'] = myAllKeywordsMap
return {'infoType': 'submission', 'infoData': parsedResult}
if __name__ == "__main__":
parseCSVFile(fileName)