def feedback_tmc():
cfg = f.load_config()
f.pnt_notice(c.msg['console_start'], os.path.basename(__file__))
f.pnt_info(c.msg["console_loading"])
teams = f.load_tsv('data_tmc')
teams.drop_duplicates(subset=['list_team'], keep='first', inplace=True)
teams = teams['list_team']
data_tmc = f.load_tsv('data_tmc')
data_tmc.replace(cfg['tmc_chart']['find_labels'],
cfg['tmc_chart']['replace_values'],
inplace=True)
tm_cols_tmc = [col for col in data_tmc.columns if 'contribution' in col]
tm_cols_tmc.sort()
tm_cols_tmc.insert(0, 'username')
tm_cols_id = [w.replace('contribution', 'id') for w in tm_cols_tmc]
f.pnt_info(c.msg["console_tmc"])
for team in teams:
print(team)
this_data = f.filter_row('data_tmc', 'list_team', team)
team_header = this_data[tm_cols_id].values.tolist()[0]
team_header[0] = 'reviews'
team_data = this_data[tm_cols_tmc].values.tolist()
this_conf_df = DataFrame.from_records(
team_data,
columns=team_header).set_index('reviews').dropna(axis=1, how='all')
this_conf_df = this_conf_df.rename(
columns=lambda x: re.sub(' - .*', '', x)).T
this_anon_df = this_conf_df.rename(columns=lambda x: re.sub(
'u.*', cfg['tmc_chart']['anon_legend'], x))
shape = this_anon_df.shape
this_anon_df['average'] = this_anon_df.mean(axis=1)
this_conf_df['average'] = this_conf_df.mean(axis=1)
f.make_tmc_chart(this_anon_df, c.d['charts'] + team + "_tmc_anon.png")
f.make_tmc_chart(this_conf_df, c.d['charts'] + team + "_tmc_conf.png")
format_tmc_feedback(team, 'conf', shape, this_data)
format_tmc_feedback(team, 'anon', shape, this_data)
f.pnt_notice(c.msg['console_complete'], os.path.basename(__file__))
all_conf = glob.glob(c.d['pdf'] + '/*_conf.pdf')
all_conf.sort()
merger = PdfFileMerger()
for pdf in all_conf:
merger.append(pdf)
merger.write(c.f['all_conf'])
merger.close()
def load_data():
'''load and clean all the csvs in the files directory'''
cfg = f.load_config()
f.pnt_notice(c.msg['console_start'], os.path.basename(__file__))
f.pnt_info(c.msg["console_loading"])
f.make_directories(c.d)
# load in all csvs in the files dir
glob.glob('./files/*.csv')
for file in glob.glob('./files/*.csv'):
this_csv = Path(file).stem
this_rename = cfg['load_file'][this_csv]['rename']
this_required = cfg['load_file'][this_csv]['required']
this_index = cfg['load_file'][this_csv]['index']
this_expected = cfg['load_file'][this_csv]['expected']
f.load_csv(this_csv)
# if this_csv == 'data_client':
# remove first two lines of the qualtrics default csv
if this_csv == 'feedback_course':
c.df[this_csv]['tutor_name'] = c.df[this_csv]['tutor'].str.replace(' ', '_')
if this_csv == 'fields':
f.col_to_lower(this_csv, "field")
if this_rename:
f.rename_header(this_csv, this_rename)
if this_csv == 'marks':
c.df[this_csv]['marker_name'] = c.df[this_csv]['marker'].str.replace('\'', '').str.replace(' ', '_')
print(c.df[this_csv]['marker_name'])
if not cfg['feedback_type']['group']:
c.df[this_csv][['user', 'name']] = c.df[this_csv]['list_name'].str.split('\s+-\s+', expand=True)
if this_index:
f.check_for_duplicates(this_csv, this_index)
if this_required:
f.check_for_empty_cells(this_csv, this_required)
if this_expected == "crit":
f.check_for_columns(this_csv)
elif this_expected == "labels":
f.check_for_labels(this_csv)
f.save_tsv(this_csv)
# print message to console - complete!
f.pnt_notice(c.msg['console_complete'], os.path.basename(__file__))
def feedback_marks():
cfg = f.load_config()
f.pnt_notice(c.msg['console_start'], os.path.basename(__file__))
f.pnt_info(c.msg["console_loading"])
# load in tsvs of needed fields
marks_df = f.load_tsv('marks')
# create a df of just the crit and the comments
crit = f.filter_row('fields', 'field', 'crit_')
comm = f.filter_row('fields', 'field', 'comment_')
f.pnt_info(c.msg["console_creating_feedback_files"])
# create distribution charts for later
if cfg['crit_display']['graph']:
stats = f.make_crit_list(crit, marks_df)
f.make_crit_chart(crit, stats, "na")
# iterate through the marks file
for record_row in marks_df.itertuples():
# decide whether to use the list_team or list_name field
if cfg['feedback_type']['group']:
this_record = record_row.list_team
this_record_name = record_row.list_team
else:
this_record = record_row.user
this_record_name = record_row.list_name
f.pnt_console(this_record)
# create the pandoc header
with open(c.d['yaml'] + this_record + '.yaml', 'w') as out:
f.pandoc_yaml(out, this_record_name)
with open(c.d['css'] + this_record + '.css', 'w') as out:
f.pandoc_css(out, this_record_name, 'anon')
#open up a file to print to
with open(c.d['md'] + this_record + '.md', 'w') as out:
print("## " + cfg['pdf_messages']['comment_title'] + "{-}\n\n", file=out)
for loop_row in comm.itertuples():
f.print_results_header(loop_row, out)
f.print_results_text(loop_row, record_row, out)
#loop through the crit columns according to app_config
if cfg['crit_display']['text']\
or cfg['crit_display']['scale']\
or cfg['crit_display']['graph']:
# start with indicator title and notes
print("# " + cfg['pdf_messages']['indicator_title'] + "{-}\n\n", file=out)
print(cfg['pdf_messages']['indicator_note'] + "\n\n", file=out)
print(cfg['pdf_messages']['chart_note'] + "\n\n", file=out)
for loop_row in crit.itertuples():
if cfg['crit_display']['text'] \
or cfg['crit_display']['scale'] \
or cfg['crit_display']['graph']:
f.print_results_header(loop_row, out)
if cfg['crit_display']['text']:
f.print_results_text(loop_row, record_row, out)
if cfg['crit_display']['scale']:
f.print_results_scale(loop_row, record_row, out)
if cfg['crit_display']['graph']:
f.print_results_graph(loop_row, record_row, out)
# if cfg['crit_display']['rubric_new_page']:
# f.print_new_page(out)
if cfg['crit_display']['rubric']:
if cfg['rubric_display']['rubric_new_page']:
print("# " + cfg['pdf_messages']['rubric_title'] + "{-}\n\n", file=out)
else:
print("## " + cfg['pdf_messages']['rubric_title'] + "{-}\n\n", file=out)
print(cfg['pdf_messages']['rubric_note'] + "\n", file=out)
f.print_results_rubric(record_row, this_record)
print("\n", file=out)
f.pandoc_html_single(this_record)
# add the rubric
if cfg['crit_display']['rubric']:
files = [c.d['rubric'] + this_record + ".html"]
with open(c.d['html'] + this_record + '.html', 'a') as outfile:
for fname in files:
with open(fname) as infile:
outfile.write(infile.read())
f.pandoc_pdf(this_record)
# print message to console - complete!
f.pnt_notice(c.msg['console_complete'], os.path.basename(__file__))
def feedback_many_eyes():
cfg = f.load_config()
# print message to console - starting!
f.pnt_notice(c.msg['console_start'], os.path.basename(__file__))
# print message to console
f.pnt_info(c.msg["console_loading"])
# load in tsvs of needed fields
fields = f.load_tsv('fields')
crit_levels = f.load_tsv('crit_levels')
data_self = f.load_tsv('data_self')
data_shadow = f.load_tsv('data_shadow')
data_tutor = f.load_tsv('data_tutor')
data_client = f.load_tsv('data_client')
# create a df of just the crit and the comments
crit = f.filter_row('fields', 'field', 'crit_')
comm = f.filter_row('fields', 'field', 'comment_')
# print message to console
f.pnt_info(c.msg["console_creating_feedback_files"])
# load data and create list of teams
teams = f.load_tsv('students')
teams.drop_duplicates(subset=['group'], keep='first', inplace=True)
# create a list of crit for access
crit_list = []
for j, row in crit.iterrows():
crit_list.append(row['label'])
# create a df for access for self and shadow
f.many_eyes_dataframe_sort('data_self')
f.many_eyes_dataframe_sort('data_shadow')
self_df = f.load_tsv('data_self_sorted')
shadow_df = f.load_tsv('data_shadow_sorted')
# create a df to access for tutor comments
tutor_df = c.df['data_tutor']
tutor_df.replace(cfg['audit_chart']['find_labels'],
cfg['audit_chart']['replace_values'],
inplace=True)
# create a df to access for client comments
client_df = c.df['data_client']
client_df.replace(cfg['audit_chart']['find_client_labels'],
cfg['audit_chart']['replace_client_values'],
inplace=True)
# work through the many eyes data
for eyes in cfg['many_eyes']['eyes']:
this_array = 'class_' + eyes + '_ave'
vars()[this_array] = []
# start a list for each eye
class_self_ave = []
class_shadow_ave = []
class_tutor_ave = []
class_client_ave = []
#loop through the crit columns
for j, row in crit.iterrows():
class_self_crit_df = self_df[self_df['crit_text'].str.contains(
row['label'])]
class_shadow_crit_df = shadow_df[shadow_df['crit_text'].str.contains(
row['label'])]
class_self_ave.append(class_self_crit_df['crit_val'].mean())
class_shadow_ave.append(class_shadow_crit_df['crit_val'].mean())
class_tutor_ave.append(tutor_df[row['field']].mean())
try:
class_client_ave.append(client_df[row['field']].mean())
except:
print("there was a key error")
class_client_ave.append(0)
# build a df for class averages
class_ave_df = pd.DataFrame()
class_ave_df['criterion'] = crit_list
class_ave_df['self'] = class_self_ave
class_ave_df['shadow'] = class_shadow_ave
class_ave_df['tutor'] = class_tutor_ave
class_ave_df['client'] = class_client_ave
class_ave_df['class_ave'] = class_ave_df.mean(axis=1)
class_ave_list = class_ave_df["class_ave"].values
# create a team list to iterate through
team_list = []
for i, row in teams.iterrows():
this_team = row['group']
team_list.append(this_team)
f.pnt_info(c.msg["console_many_eyes"])
# for each team
for team in team_list:
# report to the console
print(team)
# create a df for summary stats
this_team_self_df = self_df[self_df['team'].str.contains(team)]
this_team_shadow_df = shadow_df[shadow_df['team'].str.contains(team)]
this_team_tutor_df = tutor_df[tutor_df['team'].str.contains(team)]
this_team_client_df = client_df[client_df['team'].str.contains(
team, na=False)]
# create lists for the averages
crit_self_ave = []
crit_shadow_ave = []
crit_tutor_ave = []
crit_client_ave = []
#loop through the crit columns
for j, row in crit.iterrows():
this_team_self_crit_df = this_team_self_df[
this_team_self_df['crit_text'].str.contains(row['label'])]
this_team_shadow_crit_df = this_team_shadow_df[
this_team_shadow_df['crit_text'].str.contains(row['label'])]
# add the mean to the list
crit_self_ave.append(this_team_self_crit_df['crit_val'].mean())
crit_shadow_ave.append(this_team_shadow_crit_df['crit_val'].mean())
crit_tutor_ave.append(this_team_tutor_df[row['field']].mean())
# deal with empty client df entries
if this_team_client_df.empty:
for l, row in crit.iterrows():
crit_client_ave.append('0')
else:
for l, row in crit.iterrows():
try:
crit_client_ave.append(
this_team_client_df[row['field']].mean())
except KeyError:
crit_client_ave.append(0)
# create a df for the team to generate a graph
this_team_ave_df = pd.DataFrame()
this_team_ave_df['criterion'] = crit_list
this_team_ave_df['self'] = crit_self_ave
this_team_ave_df['shadow'] = crit_shadow_ave
this_team_ave_df['tutor'] = crit_tutor_ave
this_team_ave_df['client'] = crit_client_ave
# report the team ave and the class ave
this_team_ave_df['team_ave'] = this_team_ave_df.mean(axis=1)
this_team_ave_df['class_ave'] = class_ave_list
# now create the team chart
this_team_ave_df.set_index("criterion", drop=True, inplace=True)
f.make_audit_chart(this_team_ave_df,
c.d['charts'] + team + "_audit.png")
# run the format audit feedback for anon/conf as defined below
format_audit_feedback(team, 'conf')
format_audit_feedback(team, 'anon')
# compile the anon feedback for students
files = [
c.d['files'] + 'text_preamble.md',
c.d['md'] + team + "_tmc_anon.md",
c.d['md'] + team + "_audit_anon.md",
c.d['files'] + 'text_changelog.md'
]
with open(
c.d['md'] + team + "_" +
cfg['assignment']['assignment_short'] + "_audit_anon.md",
'w') as outfile:
for fname in files:
with open(fname) as infile:
outfile.write(infile.read())
# compile the conf feedback for teaching team
files = [
c.d['md'] + team + "_tmc_conf.md",
c.d['md'] + team + "_audit_conf.md"
]
with open(
c.d['md'] + team + "_" +
cfg['assignment']['assignment_short'] + "_audit_conf.md",
'w') as outfile:
for fname in files:
with open(fname) as infile:
outfile.write(infile.read())
# convert md to pdf
f.pandoc_html_toc(
team + "_" + cfg['assignment']['assignment_short'] + "_audit_anon",
team, 'anon')
f.pandoc_html_toc(
team + "_" + cfg['assignment']['assignment_short'] + "_audit_conf",
team, 'conf')
f.pandoc_pdf(team + "_" + cfg['assignment']['assignment_short'] +
"_audit_anon")
f.pandoc_pdf(team + "_" + cfg['assignment']['assignment_short'] +
"_audit_conf")
for k, c_row in this_team_tutor_df.iterrows():
this_text = str(c_row[this_field])
print("**Tutor**\n\n" + this_text + "\n\n", file=out)
for k, c_row in this_team_client_df.iterrows():
this_text = str(c_row[this_field])
print("**Client**\n\n" + this_text + "\n\n", file=out)
# print the confidential comments
if kind == 'conf':
print("\n\n# Confidential feedback about the team progress",
file=out)
self_conf_df = c.df['data_self'].copy()
this_team_self_conf_df = self_conf_df[self_df['team'].str.contains(
team)]
for j, row in this_team_self_conf_df.iterrows():
if (str(row['comment_confidential']) == ""):
# print messages if the comment is empty
print("**" + row['user'] + " (" + row['username'] + ")" +
"**\n\n" + cfg['tmc_pdf']['member_no_comment'] +
"\n\n",
file=out)
else:
print("**" + row['user'] + " (" + row['username'] + ")" +
"**\n\n" + str(row['comment_confidential']) + "\n\n",
file=out)
# print message to console - complete!
f.pnt_notice(c.msg['console_complete'], os.path.basename(__file__))
def wattle_csv():
cfg = f.load_config()
# print message to console - complete!
f.pnt_notice(c.msg['console_start'], os.path.basename(__file__))
# organising the files for uploading to wattle
# print message to console
f.pnt_info(c.msg["console_wattle"])
# get the list of students
f.load_tsv('students')
f.load_tsv('marks')
if cfg['feedback_type']['tmc']:
for i, row in c.df['students'].iterrows():
user = row['user']
group = row['group']
secret = hashlib.sha1(row['user'].encode('utf-8')).hexdigest()
secret_file = user + "-" + secret + ".pdf"
# cp pdf to secret here
file_from = c.d['pdf'] + group + "_tmc_anon.pdf"
file_to = c.d['upload'] + secret_file
if path.exists(file_from):
copyfile(file_from, file_to)
comment = "<a href=\"" + cfg['assignment'][
'feedback_url'] + "/" + user + "-" + secret + ".pdf\">PDF Feedback</a>"
else:
comment = "No Team Member Contribution received from the team"
# update the df
c.df['students'].at[i, 'secret'] = comment
wattle_out = c.df['students'][['user', 'secret']]
# decide whether to use the list_team or list_name field
elif not cfg['feedback_type']['group']:
# print message to console - creating secrets
f.pnt_info(c.msg['console_secrets'])
# loop through each row and create a secret for each student
for i, row in c.df['marks'].iterrows():
user = row['user']
secret = hashlib.sha1(row['user'].encode('utf-8')).hexdigest()
secret_file = user + "-" + secret + ".pdf"
comment = "<a href=\"" + cfg['assignment'][
'feedback_url'] + "/" + user + "-" + secret + ".pdf\">PDF Feedback</a>"
# update the df
c.df['marks'].at[i, 'secret'] = comment
# cp pdf to secret here
file_from = c.d['pdf'] + user + ".pdf"
file_to = c.d['upload'] + secret_file
copyfile(file_from, file_to)
marks_out = c.df['marks'][['user', 'grade_final', 'secret']]
wattle_out = marks_out.merge(c.df['students'], on='user', how='left')[[
'user', 'grade_final', 'secret'
]]
else: # use the list_team
# loop through each row and create a secret for each student
for i, row in c.df['marks'].iterrows():
group = row['list_team']
if 'suggested_indicator' in c.df['marks']:
comment = "Team Progress Indicator: " + row[
'suggested_indicator'] + "<br /><a href=\"" + cfg['assignment'][
'feedback_url'] + "/" + group + ".pdf\">PDF Feedback for " + group + "</a>"
else:
comment = "<a href=\"" + cfg['assignment'][
'feedback_url'] + "/" + group + ".pdf\">PDF Feedback for " + group + "</a>"
c.df['marks'].at[i, 'secret'] = comment
# cp pdf to secret here
file_from = c.d['pdf'] + group + ".pdf"
file_to = c.d['upload'] + group + ".pdf"
copyfile(file_from, file_to)
marks_out = c.df['marks'][['list_team', 'grade_final', 'secret']]
print(marks_out)
wattle_out = marks_out.merge(
c.df['students'],
left_on='list_team',
right_on='group',
how='left')[['user', 'grade_final', 'secret', 'group']]
print(wattle_out)
f.pnt_info(c.msg['console_upload'])
wattle_out.to_csv(c.f['wattle'], index=False)
f.pnt_notice(c.msg['console_complete'], os.path.basename(__file__))
def feedback_course():
cfg = f.load_config()
f.pnt_notice(c.msg['console_start'], os.path.basename(__file__))
f.pnt_info(c.msg["console_loading"])
tutors = f.load_tsv('feedback_course')
tutors.drop_duplicates(subset=['tutor_name'], keep='first', inplace=True)
feedback_course_df = f.load_tsv('feedback_course')
crit_levels = f.load_tsv('crit_levels')
crit = f.filter_row('fields', 'field', 'crit_')
comm = f.filter_row('fields', 'field', 'comment_')
tutor_list = f.create_list(tutors, 'tutor_name')
crit_levels_list = f.create_list(crit_levels, 'index')
f.pnt_info(c.msg["console_creating_feedback_files"])
for i, row in crit.iterrows():
this_crit = row['field']
this_crit_list = []
for tutor in tutor_list:
this_tutor_df = feedback_course_df[
feedback_course_df['tutor_name'].str.contains(tutor)]
this_tutor_crit = []
this_header = [tutor]
for val in crit_levels_list:
this_sum = (this_tutor_df[this_crit] == val).sum()
this_tutor_crit.append(this_sum)
this_crit_list.append(this_tutor_crit)
this_crit_this_tutor = pd.DataFrame(this_tutor_crit,
columns=this_header,
index=crit_levels_list)
f.make_feedback_chart(
this_crit_this_tutor,
c.d['charts'] + this_crit + "_" + tutor + ".png")
this_crit_all_tutors = pd.DataFrame(this_crit_list,
columns=crit_levels_list,
index=tutor_list)
this_crit_all_tutors = this_crit_all_tutors.T
f.make_feedback_chart(this_crit_all_tutors,
c.d['charts'] + this_crit + "_all.png")
with open(c.d['yaml'] + 'all.yaml', 'w') as out:
f.pandoc_yaml(out, 'All Tutors')
with open(c.d['css'] + 'all_conf.css', 'w') as out:
f.pandoc_css(out, 'Course Feedback', 'conf')
# open up a file to print to
with open(c.d['md'] + 'all_charts.md', 'w') as out:
print("# Quantitative Feedback\n\n", file=out)
for i, row in crit.iterrows():
this_crit = str(row['field'])
this_text = str(row['description'])
this_image = c.d['charts'] + this_crit + "_all.png"
print("### " + this_text + "\n\n", file=out)
print("![](../../." + this_image + ")\n\n", file=out)
confidential_files = [c.d['md'] + 'all_charts.md']
for tutor in tutor_list:
print(tutor)
this_tutor_df = feedback_course_df[
feedback_course_df['tutor_name'].str.contains(tutor)]
with open(c.d['yaml'] + tutor + '.yaml', 'w') as out:
f.pandoc_yaml(out, tutor)
with open(c.d['css'] + tutor + '.css', 'w') as out:
f.pandoc_css(out, tutor, 'anon')
f_out = open(c.d['md'] + tutor + '.md', 'w')
with f_out as out:
try:
for i, row in crit.iterrows():
this_crit = str(row['field'])
this_text = str(row['description'])
this_image = c.d[
'charts'] + this_crit + "_" + tutor + ".png"
print("### " + this_text + "\n", file=out)
print("![](../../." + this_image + ")\n", file=out)
for i, row in comm.iterrows():
this_field = str(row['field'])
this_description = str(row['description'])
this_df = this_tutor_df[['tutor_name',
this_field]].dropna()
if this_field != 'comment_confidential':
print("\n\n## " + this_description + "\n\n", file=out)
for i, i_row in this_df.iterrows():
this_text = str(i_row[this_field])
if (this_text != "" or this_text != "nan"
or this_text != "N/A"):
this_text_clean = BeautifulSoup(
this_text, features="html5lib")
print("**Student Comment**\n\n" +
this_text_clean.get_text() + "\n\n",
file=out)
finally:
f_out.close()
f.pandoc_html_single(tutor)
f.pandoc_pdf(tutor)
with open(c.d['md'] + tutor + '_conf.md', 'w') as out:
print("\n\n# Feedback for " + tutor + "\n\n", file=out)
confidential_files.append(c.d['md'] + tutor + '_conf.md')
# loop through the comment columns
for i, row in comm.iterrows():
this_field = str(row['field'])
this_description = str(row['description'])
this_df = this_tutor_df[['tutor_name', 'user',
this_field]].dropna()
print("\n\n## " + this_description + "\n\n", file=out)
for i, i_row in this_df.iterrows():
this_text = str(i_row[this_field])
this_user = str(i_row['user'])
if (this_text != "" or this_text != "nan"
or this_text != "N/A"):
this_text_clean = BeautifulSoup(this_text,
features="html5lib")
print("**" + this_user + "**\n\n" +
this_text_clean.get_text() + "\n\n",
file=out)
with open(c.d['md'] + "all.md", 'w') as outfile:
for fname in confidential_files:
with open(fname) as infile:
outfile.write(infile.read())
f.pandoc_html_toc('all', 'all', 'conf')
f.pandoc_pdf('all')
f.pnt_notice(c.msg['console_complete'], os.path.basename(__file__))
def analysis_many_eyes():
cfg = f.load_config()
# print message to console - starting!
f.pnt_notice(c.msg['console_start'], os.path.basename(__file__))
# print message to console
f.pnt_info(c.msg["console_loading"])
# load in tsvs of needed fields
student = f.load_tsv('students')
student.drop_duplicates(keep='first', inplace=True)
student.dropna(how='any', subset=['user'], inplace=True)
students = student.copy()
fields = f.load_tsv('fields')
crit_levels = f.load_tsv('crit_levels')
data_self = f.load_tsv('data_self')
data_shadow = f.load_tsv('data_shadow')
data_tutor = f.load_tsv('data_tutor')
data_client = f.load_tsv('data_client')
# create a df of just the crit and the comments
crit = f.filter_row('fields', 'field', 'crit_')
comm = f.filter_row('fields', 'field', 'comment_')
# print message to console
f.pnt_info(c.msg["console_reading_feedback_files"])
# load data and create list of teams
teams = f.load_tsv('students')
teams.drop_duplicates(subset=['group'], keep='first', inplace=True)
# create a team list to iterate through
team_list = []
for i, row in teams.iterrows():
this_team = row['group']
team_list.append(this_team)
# create a list of crit for access
crit_list = []
crit_list_header = ['team', 'role']
for j, row in crit.iterrows():
crit_list.append(row['label'])
crit_list_header.append(row['label'])
# create a list of comments for access
comment_list = f.create_list(comm, 'field')
# create a df for access for self and shadow
f.many_eyes_dataframe_sort('data_self')
f.many_eyes_dataframe_sort('data_shadow')
self_df = f.load_tsv('data_self_sorted')
shadow_df = f.load_tsv('data_shadow_sorted')
# create a df to access for tutor comments
tutor_df = c.df['data_tutor']
tutor_df.replace(cfg['audit_chart']['find_labels'],
cfg['audit_chart']['replace_values'],
inplace=True)
# create a df to access for client comments
client_df = c.df['data_client']
client_df.replace(cfg['audit_chart']['find_client_labels'],
cfg['audit_chart']['replace_client_values'],
inplace=True)
f.pnt_info(c.msg["console_many_eyes"])
print("Cleaning the self data...")
for i, row in self_df.iterrows():
for readability_list in cfg['analytics']['readability_stats']:
f.readability_stats('data_self_sorted', row, i, 'crit_comment',
readability_list[0], readability_list[1],
readability_list[2])
print("Cleaning the shadow data...")
for i, row in shadow_df.iterrows():
for readability_list in cfg['analytics']['readability_stats']:
f.readability_stats('data_shadow_sorted', row, i, 'crit_comment',
readability_list[0], readability_list[1],
readability_list[2])
team_stats_list = []
# report to the console
print('Generating team stats...')
# for each team
for team in team_list:
# work through the many eyes data
for eye in cfg['many_eyes']['eyes']:
this_stats_list = [team, eye]
this_df = locals()[eye + "_df"]
this_crit_df = this_df[this_df['team'].str.contains(team,
na=False)]
if not this_crit_df.empty:
for j, row in crit.iterrows():
if eye == 'self' or eye == 'shadow':
this_crit_df = this_df[
this_df['crit_text'].str.contains(row['label'])]
this_stats_list.append(this_crit_df['crit_val'].mean())
else:
try:
this_stats_list.append(
this_crit_df[row['field']].mean())
except:
this_stats_list.append(0)
team_stats_list.append(this_stats_list)
team_stats = pd.DataFrame(team_stats_list, columns=crit_list_header)
team_stats_ave_list = []
for team in team_list:
this_team_stats = team_stats[team_stats['team'].str.contains(team)]
this_team_ave_list = [team, 'average']
for crit in crit_list:
this_team_crit_mean = this_team_stats[crit].mean()
this_team_ave_list.append(this_team_crit_mean)
team_stats_ave_list.append(this_team_ave_list)
team_stats_ave_df = pd.DataFrame(team_stats_ave_list,
columns=crit_list_header)
frames = [team_stats, team_stats_ave_df]
team_stats_df = pd.concat(frames)
crit = f.filter_row('fields', 'field', 'crit_')
# report to the console
print('Generating team charts...')
for team in team_list:
print(team)
#loop through the crit columns
for i, row in crit.iterrows():
this_crit_label = row['label']
this_crit_field = row['field']
this_team_crit_list = []
this_team_crit_header = ['role', 'value']
for eye in cfg['many_eyes']['eyes']:
this_df = locals()[eye + "_df"]
this_team_df = this_df[this_df['team'].str.contains(team,
na=False)]
if not this_team_df.empty:
if eye == 'self' or eye == 'shadow':
this_team_crit_df = this_team_df[this_team_df[
'crit_text'].str.contains(this_crit_label)]
for j, row in this_team_crit_df.iterrows():
this_team_crit_list.append([eye, row['crit_val']])
else:
for j, row in this_team_df.iterrows():
try:
this_team_crit_list.append(
[eye, row[this_crit_field]])
except:
this_team_crit_list.append([eye, 0])
this_team_crit_df = pd.DataFrame(this_team_crit_list,
columns=this_team_crit_header).T
this_team_crit_df.columns = this_team_crit_df.iloc[0]
this_team_crit_df = this_team_crit_df.iloc[1:].rename_axis(None,
axis=1)
this_team_crit_df['average'] = this_team_crit_df.mean(axis=1)
f.make_audit_crit_chart(
this_team_crit_df,
c.d['charts'] + team + "_" + this_crit_field + "_audit.png")
self_crit_val_diff_list = []
self_crit_val_abs_list = []
self_wps_list = []
for i, row in self_df.iterrows():
this_crit_val = row['crit_val']
this_crit_text = row['crit_text']
this_team = row['team']
this_wc = row['wc']
this_sc = row['sc']
this_team_row = team_stats_ave_df[
team_stats_ave_df['team'].str.contains(this_team)]
# this_crit_ave = this_team_row.iloc[0][this_crit_text]
this_crit_ave = 0
this_diff = this_crit_val - this_crit_ave
this_diff_abs = abs(this_diff)
self_crit_val_diff_list.append(this_diff)
self_crit_val_abs_list.append(this_diff_abs)
this_wps = this_wc / this_sc
self_wps_list.append(this_wps)
self_df['crit_val_diff'] = self_crit_val_diff_list
self_df['crit_val_abs'] = self_crit_val_abs_list
self_df['wps'] = self_wps_list
shadow_crit_val_diff_list = []
shadow_crit_val_abs_list = []
shadow_wps_list = []
for i, row in shadow_df.iterrows():
print(shadow_df)
this_crit_val = row['crit_val']
this_crit_text = row['crit_text']
this_team = row['team']
this_wc = row['wc']
this_sc = row['sc']
this_team_row = team_stats_ave_df[
team_stats_ave_df['team'].str.contains(this_team)]
if not this_team_row.empty:
this_crit_ave = this_team_row.iloc[0][this_crit_text]
else:
this_crit_ave = 0
this_diff = this_crit_val - this_crit_ave
this_diff_abs = abs(this_diff)
shadow_crit_val_diff_list.append(this_diff)
shadow_crit_val_abs_list.append(this_diff_abs)
this_wps = this_wc / this_sc
shadow_wps_list.append(this_wps)
shadow_df['crit_val_diff'] = shadow_crit_val_diff_list
shadow_df['crit_val_abs'] = shadow_crit_val_abs_list
shadow_df['wps'] = shadow_wps_list
self_df['wc_pct_rank'] = self_df.wc.rank(pct=True)
self_df['wps_pct_rank'] = self_df.sc.rank(pct=True)
self_df['flesch_pct_rank'] = self_df.flesch.rank(pct=True)
self_df['crit_val_diff_rank'] = self_df.crit_val_diff.rank(pct=True)
shadow_df['wc_pct_rank'] = shadow_df.wc.rank(pct=True)
shadow_df['wps_pct_rank'] = shadow_df.sc.rank(pct=True)
shadow_df['flesch_pct_rank'] = shadow_df.flesch.rank(pct=True)
shadow_df['crit_val_diff_rank'] = shadow_df.crit_val_diff.rank(pct=True)
wc_score = [-2, -1, 0, 1, 2]
wps_score = [-2, -1, 0, 1, 2]
flesch_score = [2, 1, 0, -1, -2]
cvd_score = [2, 1, 0, -1, -2]
wc_score_actual = [-2, -1, 0, 1, 2]
score_pct = [0.2, 0.4, 0.6, 0.8, 1.0]
score_act = [25, 50, 100, 200, 10000]
score_dataframe(self_df, 'wc_pct_rank', 'wc_score', wc_score, score_pct)
score_dataframe(shadow_df, 'wc_pct_rank', 'wc_score', wc_score, score_pct)
score_dataframe(self_df, 'wps_pct_rank', 'wps_score', wps_score, score_pct)
score_dataframe(shadow_df, 'wps_pct_rank', 'wps_score', wps_score,
score_pct)
score_dataframe(self_df, 'flesch_pct_rank', 'flesch_score', flesch_score,
score_pct)
score_dataframe(shadow_df, 'flesch_pct_rank', 'flesch_score', flesch_score,
score_pct)
score_dataframe(self_df, 'crit_val_diff_rank', 'crit_val_score', cvd_score,
score_pct)
score_dataframe(shadow_df, 'crit_val_diff_rank', 'crit_val_score',
cvd_score, score_pct)
score_dataframe(self_df, 'wc', 'wc_score_act', wc_score, score_act)
score_dataframe(shadow_df, 'wc', 'wc_score_act', wc_score, score_act)
col_chart_list = ["wc", "wps", "flesch", "crit_val_diff"]
for col in col_chart_list:
if col == "flesch":
chart_min = 0
chart_max = 100
elif col == "crit_val_diff":
chart_min = -5
chart_max = 5
else:
chart_min = 0
self_max = self_df[col].max()
shadow_max = shadow_df[col].max()
if self_max > shadow_max:
this_max = self_max
else:
this_max = shadow_max
chart_max = int(round(this_max / 20)) * 20
f.make_col_chart(self_df, col, 'self', chart_min, chart_max)
f.make_col_chart(shadow_df, col, 'shadow', chart_min, chart_max)
# for each team
for i, row in students.iterrows():
# report to the console
this_user = str(row['user'])
this_first = str(row['firstname'])
this_last = str(row['lastname'])
self_team = str(row['group'])
shadow_team = str(row['shadowteam'])
print(this_user)
this_self_df = self_df[self_df['username'].str.contains(this_user)]
this_shadow_df = shadow_df[shadow_df['username'].str.contains(
this_user)]
if this_self_df.empty and this_shadow_df.empty:
with open(c.d["md"] + str(this_user) + '.md', 'w') as out:
print("## " + this_user + " - " + this_first + " " +
this_last + "\n\n",
file=out)
print("**No reviews received.**\n\n", file=out)
else:
try:
local_text_analysis(this_self_df, this_user, 'self')
local_text_analysis(this_shadow_df, this_user, 'shadow')
with open(c.d["md"] + str(this_user) + '.md', 'w') as out:
print("## " + this_user + " - " + this_first + " " +
this_last + "\n\n",
file=out)
print(
"**This feedback report shows analytics on the feedback you gave in the review.**\n\n",
file=out)
print(
"By considering the analytics in this report, you should get a sense of how your feedback compares to others.\n\n",
file=out)
print("# Project Audits\n\n", file=out)
print("*Compare your review to others' perspectives.*\n\n",
file=out)
print("## Project Team: " + self_team + "\n\n", file=out)
if this_self_df.empty:
print("* No review submitted", file=out)
else:
for i, row in this_self_df.iterrows():
print("*Your review of " + str(row['crit_text']) +
":* **" + str(row['crit_desc']) + "**\n\n",
file=out)
this_crit = f.filter_row('fields', 'label',
row['crit_text'])
for j, j_row in this_crit.iterrows():
this_crit_field = j_row['field']
print("![](../../." + c.d['charts'] +
self_team + "_" + this_crit_field +
"_audit.png)\n\n",
file=out)
print("\n*Chart of all reviews of " + self_team +
" for " + str(row['crit_text']) + "*\n\n",
file=out)
print("<div class=\"new-page\">.</div>", file=out)
print("## Shadow Team: " + shadow_team + "\n\n", file=out)
if this_shadow_df.empty:
print("* No review submitted", file=out)
else:
for i, row in this_shadow_df.iterrows():
print("*Your review of " + str(row['crit_text']) +
":* **" + str(row['crit_desc']) + "**\n\n",
file=out)
this_crit = f.filter_row('fields', 'label',
row['crit_text'])
for j, j_row in this_crit.iterrows():
this_crit_field = j_row['field']
print("![](../../." + c.d['charts'] +
shadow_team + "_" + this_crit_field +
"_audit.png)\n\n",
file=out)
print("\n*Chart of all reviews of " + shadow_team +
" for " + str(row['crit_text']) + "*\n\n",
file=out)
print("<div class=\"new-page\">.</div>", file=out)
print("## Difference in review evaluations\n\n", file=out)
print(
"*A negative value means that your review was below the average of reviews; positive, above. '2' represents a 'full band', such as the difference between 'Baseline' and 'Acceptable'.*\n\n",
file=out)
this_head = ["Team", "Role"]
this_self_list = [self_team, "Self"]
this_shadow_list = [shadow_team, "Shadow"]
if not this_self_df.empty:
for i, row in this_self_df.iterrows():
try:
this_head.append(str(row['crit_text']))
this_self_list.append("**" + str(
int(round(row['crit_val_diff'] * 4) / 4)) +
"** half-bands")
except:
this_self_list.append("**N/A**")
this_head.append("Class Average")
try:
this_self_list.append(
str(
int(
round(this_self_df['crit_val_abs'].mean() *
4) / 4)) + " (self)")
except:
this_self_list.append("**N/A**")
if not this_shadow_df.empty:
for i, row in this_shadow_df.iterrows():
try:
this_head.append(str(row['crit_text']))
this_shadow_list.append("**" + str(
int(round(row['crit_val_abs'] * 4) / 4)) +
"** half-bands")
except:
this_self_list.append("**N/A**")
this_head.append("Class Average")
try:
this_shadow_list.append(
str(
int(
round(this_shadow_df['crit_val_abs'].mean(
) * 4) / 4)) + " (shadow)")
except:
this_shadow_list.append("**N/A**")
this_html_df = pd.DataFrame(
list(zip(this_head, this_self_list,
this_shadow_list))).T
this_html_df.columns = this_html_df.iloc[0]
this_html_df = this_html_df.iloc[1:].rename_axis(None,
axis=1)
# this_html_df.set_index("Team",drop=True,inplace=True)
print(this_html_df.to_html(), file=out)
print("![](../../." + c.d['charts'] +
"crit_val_diff_self.png)\n",
file=out)
print("![](../../." + c.d['charts'] +
"crit_val_diff_shadow.png)\n",
file=out)
print(
"\n*Top: Histogram of differences for self reviews.*",
file=out)
print(
"*Bottom: Histogram of differences for shadow reviews.*\n\n",
file=out)
print("<div class=\"new-page\">.</div>", file=out)
print("# Descriptive statistics about your comments\n\n",
file=out)
print("## Word Count\n\n", file=out)
print(
"The course guide requests you to complete 250-500 words per review.\n\n",
file=out)
print("*Table of your review word count statistics*\n\n",
file=out)
this_head = ["Team", "Role"]
this_self_list = [self_team, "Self"]
this_shadow_list = [shadow_team, "Shadow"]
if not this_self_df.empty:
for i, row in this_self_df.iterrows():
try:
this_head.append(str(row['crit_text']))
this_self_list.append(
"**" + str(int(round(row['wc']))) +
"** words")
except:
this_self_list.append("**N/A**")
this_head.append("Class Average")
try:
this_self_list.append(
str(int(round(this_self_df['wc'].mean()))) +
" (self)")
except:
this_self_list.append("**N/A**")
if not this_shadow_df.empty:
for i, row in this_shadow_df.iterrows():
try:
this_head.append(str(row['crit_text']))
this_shadow_list.append(
"**" + str(int(round(row['wc']))) +
"** words")
except:
this_shadow_list.append("**N/A**")
this_head.append("Class Average")
try:
this_shadow_list.append(
str(int(round(this_shadow_df['wc'].mean()))) +
" (shadow)")
except:
this_shadow_list.append("**N/A**")
this_html_df = pd.DataFrame(
list(zip(this_head, this_self_list,
this_shadow_list))).T
this_html_df.columns = this_html_df.iloc[0]
this_html_df = this_html_df.iloc[1:].rename_axis(None,
axis=1)
# this_html_df.set_index("Team",drop=True,inplace=True)
print(this_html_df.to_html(), file=out)
print("![](../../." + c.d['charts'] + "wc_self.png)\n",
file=out)
print("![](../../." + c.d['charts'] + "wc_shadow.png)\n",
file=out)
print(
"\n*Top: Histogram of word counts for self reviews.*",
file=out)
print(
"*Bottom: Histogram of word counts for shadow reviews.*\n\n",
file=out)
print("<div class=\"new-page\">.</div>", file=out)
print("## Words per Sentence Count\n\n", file=out)
print(
"Generally, shorter sentences are easier to read.\n\n",
file=out)
print(
"*Table of your review words per sentence count statistics*\n\n",
file=out)
this_head = ["Team", "Role"]
this_self_list = [self_team, "Self"]
this_shadow_list = [shadow_team, "Shadow"]
if not this_self_df.empty:
for i, row in this_self_df.iterrows():
try:
this_head.append(str(row['crit_text']))
this_self_list.append(
"**" + str(int(round(row['wps']))) +
"** wps")
except:
this_self_list.append("**N/A**")
this_head.append("Class Average")
try:
this_self_list.append(
str(int(round(this_self_df['wps'].mean()))) +
" (self)")
except:
this_self_list.append("**N/A**")
if not this_shadow_df.empty:
for i, row in this_shadow_df.iterrows():
try:
this_head.append(str(row['crit_text']))
this_shadow_list.append(
"**" + str(int(round(row['wps']))) +
"** wps")
except:
this_shadow_list.append("**N/A**")
this_head.append("Class Average")
try:
this_shadow_list.append(
str(int(round(this_shadow_df['wps'].mean()))) +
" (shadow)")
except:
this_shadow_list.append("**N/A**")
this_html_df = pd.DataFrame(
list(zip(this_head, this_self_list,
this_shadow_list))).T
this_html_df.columns = this_html_df.iloc[0]
this_html_df = this_html_df.iloc[1:].rename_axis(None,
axis=1)
# this_html_df.set_index("Team",drop=True,inplace=True)
print(this_html_df.to_html(), file=out)
print("![](../../." + c.d['charts'] + "wps_self.png)\n",
file=out)
print("![](../../." + c.d['charts'] + "wps_shadow.png)\n",
file=out)
print(
"\n*Top: Histogram of sentence counts for self reviews.*",
file=out)
print(
"*Bottom: Histogram of sentence counts for shadow reviews.*\n\n",
file=out)
print("<div class=\"new-page\">.</div>", file=out)
print("## Flesch–Kincaid readability test\n\n", file=out)
print(
"Based on a score out of 100, lower scores are typically harder to read: scores below 50 are considered difficult to read. [More information on Wikipedia](https://en.wikipedia.org/wiki/Flesch%E2%80%93Kincaid_readability_tests)\n\n",
file=out)
print(
"*Table of your Flesch–Kincaid readability score statistics*\n\n",
file=out)
this_head = ["Team", "Role"]
this_self_list = [self_team, "Self"]
this_shadow_list = [shadow_team, "Shadow"]
if not this_self_df.empty:
for i, row in this_self_df.iterrows():
try:
this_head.append(str(row['crit_text']))
this_self_list.append(
"**" + str(int(round(row['flesch']))) +
"**")
except:
this_self_list.append("**N/A**")
this_head.append("Class Average")
try:
this_self_list.append(
str(int(round(this_self_df['flesch'].mean()))) +
" (self)")
except:
this_self_list.append("**N/A**")
if not this_shadow_df.empty:
for i, row in this_shadow_df.iterrows():
try:
this_head.append(str(row['crit_text']))
this_shadow_list.append(
"**" + str(int(round(row['flesch']))) +
"**")
except:
this_shadow_list.append("**N/A**")
this_head.append("Class Average")
try:
this_shadow_list.append(
str(int(round(this_shadow_df['flesch'].mean()))) +
" (shadow)")
except:
this_shadow_list.append("**N/A**")
this_html_df = pd.DataFrame(
list(zip(this_head, this_self_list,
this_shadow_list))).T
this_html_df.columns = this_html_df.iloc[0]
this_html_df = this_html_df.iloc[1:].rename_axis(None,
axis=1)
# this_html_df.set_index("Team",drop=True,inplace=True)
print(this_html_df.to_html(), file=out)
print("![](../../." + c.d['charts'] + "flesch_self.png)\n",
file=out)
print("![](../../." + c.d['charts'] +
"flesch_shadow.png)\n",
file=out)
print(
"\n*Top: Histogram of Flesch-Kincaid readability for self reviews.*",
file=out)
print(
"*Bottom: Histogram of Flesch-Kincaid readability for shadow reviews.*\n\n",
file=out)
role_list = ["self", "shadow"]
print("# Analytics generated from your comments", file=out)
print("\n*" + cfg['analytics']['nlp_source_comment'] +
"*\n",
file=out)
# using enumerate to access list indices for name and title
# work through the defined nlp endpoints
for num, endpoint in enumerate(cfg['aylien']['endpoints'],
start=0):
name = (cfg['aylien']['endpoint_name'][num])
title = (cfg['aylien']['endpoint_title'][num])
print("<article id=\"columns\">", file=out)
print("\n## Comment " + title + "\n\n", file=out)
if endpoint == "sentiment":
print("<section>", file=out)
# loop through the analysis for each comment
for role in role_list:
# load the nlp json response to read from
with open(c.d['nlp'] + this_user + "_comment_" +
role + ".json") as json_file:
this_nlp = json.load(json_file)
# print a header to out
print("\n**" + title + " for " + role +
" comments**"
"\n\n",
file=out)
if endpoint == "sentiment":
item_out = ""
for item in this_nlp[name]:
item_out += item
print("* " + item_out, file=out)
else:
try:
item_check = []
item_out = ""
for item in this_nlp[name]:
# replace hashes so that they are not interpreted in markdown
if endpoint != 'entities':
item_out = item.replace(
"#", "\\#")
print("* " + item_out,
file=out)
else:
this_item = item.split(' ')
item_out = this_item[
0].replace("#", "\\#")
if item_out not in item_check:
# print to out each item in a list
print("* " + item_out,
file=out)
item_check.append(item_out)
except:
# if there's nothing there, print N/A
print("* N/A", file=out)
if endpoint == "sentiment":
print("</section>", file=out)
print("</article>", file=out)
print("\n\n# Wordclouds generated from your comments\n",
file=out)
# loop through the analysis for each comment
for role in role_list:
print("**Your " + role + " review**\n\n", file=out)
print("![](../../." + c.d['wordcloud'] + this_user +
"_" + role + ".png)\n",
file=out)
print(
"\n*Above: Wordcloud generated from feedback you gave in "
+ role + " reviews.*\n",
file=out)
print("\n\n# Record of submissions", file=out)
if this_self_df.empty:
print("* No review submitted", file=out)
else:
for i, row in this_self_df.iterrows():
print("**Your review of " + str(row['crit_text']) +
" for " + self_team + "**\n\n",
file=out)
print(str(row['crit_comment']) + "\n\n", file=out)
if this_shadow_df.empty:
print("* No review submitted", file=out)
else:
for i, row in this_shadow_df.iterrows():
print("**Your review of " + str(row['crit_text']) +
" for " + shadow_team + "**\n\n",
file=out)
print(str(row['crit_comment']) + "\n\n", file=out)
except:
with open(c.d["md"] + str(this_user) + '.md', 'w') as out:
print("## " + this_user + " - " + this_first + " " +
this_last + "\n\n",
file=out)
print("**Error in review compilation.**\n\n", file=out)
# create the weasyprint variables
with open(c.d['yaml'] + str(this_user) + '.yaml', 'w') as out:
f.pandoc_yaml(out, str(this_user))
# create the weasyprint stylesheet
with open(c.d['css'] + str(this_user) + ".css", 'w') as out:
f.pandoc_css(out, str(this_user), 'anon')
# turn the pandoc md to html to pdf
f.pandoc_html_single(str(this_user))
f.pandoc_pdf(str(this_user))
# loop through each row and create a secret for each student
for i, row in students.iterrows():
this_user = row['user']
secret = hashlib.sha1(row['user'].encode('utf-8')).hexdigest()
secret_file = this_user + "-" + secret + ".pdf"
comment = "<a href=\"" + cfg['assignment'][
'feedback_url'] + "reviews/" + this_user + "-" + secret + ".pdf\">PDF Feedback on your Reviews</a>"
# update the df
students.loc[i, 'secret'] = comment
# cp pdf to secret here
file_from = c.d['pdf'] + this_user + ".pdf"
file_to = c.d['review'] + secret_file
copyfile(file_from, file_to)
this_self_df = self_df[self_df['username'].str.contains(this_user)]
this_shadow_df = shadow_df[shadow_df['username'].str.contains(
this_user)]
self_wc_score = this_self_df['wc_score'].sum()
shadow_wc_score = this_shadow_df['wc_score'].sum()
self_wps_score = this_self_df['wps_score'].sum()
shadow_wps_score = this_shadow_df['wps_score'].sum()
self_flesch_score = this_self_df['flesch_score'].sum()
shadow_flesch_score = this_shadow_df['flesch_score'].sum()
self_crit_score = this_self_df['crit_val_score'].sum()
shadow_crit_score = this_shadow_df['crit_val_score'].sum()
this_text_score = self_wc_score + shadow_wc_score + self_wps_score + shadow_wps_score + self_flesch_score + shadow_flesch_score
this_crit_score = self_crit_score + shadow_crit_score
students.loc[i, 'text_score'] = this_text_score
students.loc[i, 'crit_score'] = this_crit_score
students['text_rank'] = students.text_score.rank(pct=True)
students['crit_rank'] = students.crit_score.rank(pct=True)
for i, row in students.iterrows():
this_text_rank = row['text_rank']
this_crit_rank = row['crit_rank']
this_combined_rank = this_text_rank + this_crit_rank
students.loc[i, 'combined_rank'] = this_combined_rank
students['score_rank'] = students.combined_rank.rank(pct=True)
this_out = students[[
'user', 'secret', 'text_score', 'crit_score', 'score_rank'
]]
this_out.to_csv(c.f['wattle_analysis'], index=False)
def wattle_csv_many_eyes():
cfg = f.load_config()
f.pnt_notice(c.msg['console_start'], os.path.basename(__file__))
f.pnt_info(c.msg["console_wattle"])
f.load_tsv('data_tutor')
teams = f.load_tsv('students')
teams.drop_duplicates(subset=['group'], keep='first', inplace=True)
for i, row in teams.iterrows():
this_team = row['group']
# cp pdf to secret here
file_from = c.d['pdf'] + this_team + "_" + cfg['assignment'][
'assignment_short'] + "_audit_anon.pdf"
file_to = c.d['upload'] + this_team + "_" + cfg['assignment'][
'assignment_short'] + ".pdf"
copyfile(file_from, file_to)
f.load_tsv('students')
students = f.filter_row('students', 'role', 'Student')
user_list = []
comment_list = []
for i, row in students.iterrows():
user = row['user']
role = row['role']
project_team = row['group']
team_row = f.filter_row('data_tutor', 'team', project_team)
if not team_row.empty:
team_performance = team_row.iloc[0]['suggestedindicator']
else:
team_performance = "TBA"
if role != 'student':
project_team = row['group']
shadow_team = row['shadowteam']
comment = "Your Team's Progress Indicator:"
comment += "<ul><li><strong>" + team_performance + "</strong></li></ul>"
comment += "Feedback for Your Teams:"
comment += "<ul><li><a href=\"" + cfg['assignment'][
'feedback_url'] + "/" + str(
project_team
) + "_" + cfg['assignment'][
'assignment_short'] + ".pdf\">PDF Feedback for your team: " + str(
project_team) + "</a></li>"
comment += "<li><a href=\"" + cfg['assignment'][
'feedback_url'] + "/" + str(
shadow_team
) + "_" + cfg['assignment'][
'assignment_short'] + ".pdf\">PDF Feedback for shadow team: " + str(
shadow_team) + "</a></li></ul>"
user_list.append(user)
comment_list.append(comment)
f.pnt_info(c.msg['console_upload'])
this_out = pd.DataFrame()
this_out['users'] = user_list
this_out['feedback'] = comment_list
this_out.to_csv(c.f['wattle'], index=False)
f.pnt_notice(c.msg['console_complete'], os.path.basename(__file__))
def analysis_marks():
cfg = f.load_config()
f.pnt_notice(c.msg['console_start'], os.path.basename(__file__))
f.pnt_info(c.msg["console_loading"])
marks_df = f.load_tsv('marks')
crit_levels = f.load_tsv('crit_levels')
# create a df of just the crit and the comments
crit = f.filter_row('fields', 'field', 'crit_')
comm = f.filter_row('fields', 'field', 'comment_')
crit_list = f.create_list(crit, 'field')
comment_list = f.create_list(comm, 'field')
f.pnt_info(c.msg["console_creating_feedback_files"])
# clean up marks df
c.df['marks'].loc[c.df['marks'].grade_final == 0, ['marker_name', 'marker_id']] = 'No_Submission', 'nil'
for comment in comment_list:
c.df['marks'][comment + "_txt"] = c.df['marks'][comment].apply(f.html_to_text)
f.pnt_console("Processing the marks file")
for i, row in c.df['marks'].iterrows():
for comment in comment_list:
for readability_list in cfg['analytics']['readability_stats']:
f.readability_stats('marks', row, i, comment + '_txt', comment + "_" + readability_list[0], readability_list[1], readability_list[2])
c.df['marks']['grade_final_pct'] = ( c.df['marks']['grade_final'] / int(cfg['assignment']['grade_final_out_of']) * 100 )
c.df['marks']['diff_final_sugg'] = (c.df['marks']['grade_final_pct'] - c.df['marks']['grade_suggested']).round(decimals=1)
c.df['marks']['diff_calc_sugg'] = (c.df['marks']['grade_calculated'] - c.df['marks']['grade_suggested']).round(decimals=1)
print(c.df['marks']['grade_final'])
f.make_hist_chart(c.df['marks'], 'grade_final_pct')
f.make_boxplot_chart(c.df['marks'], 'grade_final_pct')
# generate a dataframe with the readability stats for each marker
# then replace any nil submissions with 'no submission'
marker = c.df['marks']['grade_final'].groupby([c.df['marks']['marker_name']]).mean().reset_index()
marker = marker[marker.marker_name != 'No_Submission']
# create a marker_list for iteration
marker_list = f.create_list(marker, 'marker_name')
f.pnt_info("Building crit graphs...")
# work through each criterion
for criterion in crit_list:
# print to tho console
print(criterion)
this_crit_df = crit_levels['index']
for i, row in marker.iterrows():
this_marker_name = row['marker_name']
this_marker_df = f.filter_row('marks', 'marker_name', this_marker_name)
this_marker_stats = f.make_crit_list(crit, this_marker_df)
this_col_sum = this_marker_stats[criterion].sum()
this_marker_stats[this_marker_name] = this_marker_stats[criterion].apply(lambda x: x/this_col_sum*100)
this_crit_df = pd.merge(this_crit_df, this_marker_stats[this_marker_name], on='index')
print(this_crit_df)
f.make_stacked_chart(this_crit_df, criterion, True)
bin_values = [-10, -5.5, -2.5, -1.5, -0.5, 0.5, 1.5, 2.5, 5.5, 10]
bin_labels = [-10, -5, -2, -1, 0, 1, 2, 5, 10]
calc_sugg_df = pd.DataFrame(bin_labels,columns=['bin'])
final_sugg_df = pd.DataFrame(bin_labels,columns=['bin'])
c.df['marks']['bin_calc_sugg'] = pd.cut(c.df['marks']['diff_calc_sugg'], bins=bin_values, labels=bin_labels)
c.df['marks']['bin_final_sugg'] = pd.cut(c.df['marks']['diff_final_sugg'], bins=bin_values, labels=bin_labels)
for i, row in marker.iterrows():
this_marker_name = row['marker_name']
this_marker_df = f.filter_row('marks', 'marker_name', this_marker_name)
this_marker_calc_df = pd.DataFrame(this_marker_df.bin_calc_sugg.value_counts()).reset_index()
this_marker_calc_df = this_marker_calc_df.rename(columns={'index': 'bin'})
this_col_sum = this_marker_calc_df['bin_calc_sugg'].sum()
this_marker_calc_df[this_marker_name] = this_marker_calc_df['bin_calc_sugg'].apply(lambda x: x/this_col_sum*100)
this_marker_calc_df = this_marker_calc_df.drop(['bin_calc_sugg'], axis=1)
diff_calc_sugg_df = pd.merge(calc_sugg_df, this_marker_calc_df, on='bin')
calc_sugg_df = diff_calc_sugg_df.set_index("bin")
calc_sugg_df.reset_index(inplace=True)
calc_sugg_df = calc_sugg_df.rename(columns={'bin': 'index'})
print(calc_sugg_df)
f.make_stacked_chart(calc_sugg_df, 'suggested', False)
f.pnt_info("Analysing each marker...")
for i, row in marker.iterrows():
this_marker_name = row['marker_name']
print(this_marker_name)
this_group_df=f.filter_row('marks', 'marker_name', row['marker_name'])
marker.at[i, 'grade_count'] = this_group_df['grade_final'].count()
marker.at[i, 'grade_mean'] = this_group_df['grade_final'].mean()
marker.at[i, 'grade_min'] = this_group_df['grade_final'].min()
marker.at[i, 'grade_max'] = this_group_df['grade_final'].max()
marker.at[i, 'grade_std'] = this_group_df['grade_final'].std()
marker.at[i, 'grade_skew'] = this_group_df['grade_final'].skew()
# generate the readability stats
for comment in comment_list:
for readability_list in cfg['analytics']['readability_stats']:
marker.at[i, comment + "_" + readability_list[0]] = this_group_df[comment + "_" + readability_list[0]].mean()
# get the comment to pass to the nlp
this_comment = " ".join(f.create_list(this_group_df, comment + "_txt"))
# store the comment so that it can be accessed for the word cloud
with open(c.d['txt'] + this_marker_name + "_" + comment + ".txt", 'w') as out:
print(this_comment, file=out)
this_nlp_file = Path(c.d['nlp'] + this_marker_name + "_" + comment + ".json")
# check if the nlp data already exists
# this is important only to reduce the numbers of calls on the api for local testing
if not this_nlp_file.is_file():
# get the results from the api
this_nlp = f.text_analysis_api(this_comment, 'comment', row['marker_name'])
with open(this_nlp_file, 'w') as out:
print(this_nlp, file=out)
# create a wordcloud using the wordcloud_cli interface
subprocess.call("wordcloud_cli --width 1000 --height 250 --text " + c.d['txt'] + this_marker_name + "_" + comment + ".txt --imagefile " + c.d['wordcloud'] + this_marker_name + "_" + comment + ".png --fontfile ./includes/fonts/Roboto-Medium.ttf --background white --color blue", shell=True)
# open a file to write to for each marker
# this will be joined to the end of the analysis
with open(c.d['md'] + this_marker_name + '.md', 'w') as out:
print("\n\n## Analysis of " + this_marker_name + "'s Feedback\n\n", file=out)
# loop through the analysis for each comment
for i, row in comm.iterrows():
comment = row['field']
field_text = row['label']
print("\n\n## Wordcloud for " + field_text + "\n\n", file=out)
print("![](../../." + c.d['wordcloud'] + this_marker_name + "_" + comment + ".png) \n\n", file=out)
# using enumerate to access list indices for name and title
# work through the defined nlp endpoints
for num, endpoint in enumerate(cfg['aylien']['endpoints'], start=0):
name = (cfg['aylien']['endpoint_name'][num])
title = (cfg['aylien']['endpoint_title'][num])
# treat sentiment differently, as it is better presented in a table
if endpoint != 'sentiment':
with open(c.d['md'] + this_marker_name + '.md', 'a') as out:
print("\n### Comment " + title + "\n\n", file=out)
print("\n*" + cfg['analytics']['nlp_source_comment']+ "*\n", file=out)
# loop through the analysis for each comment
for i, row in comm.iterrows():
comment = row['field']
field_text = row['label']
# load the nlp json response to read from
with open(c.d['nlp'] + this_marker_name + "_" + comment + ".json") as json_file:
this_nlp = json.load(json_file)
# print a header to out
print("\n**" + title + " for " + field_text + "**" "\n\n", file=out)
try:
item_out = ""
for item in this_nlp[name]:
item_out = item.replace("#", "\\#")
print("* " + item_out, file=out)
except:
print("* N/A", file=out)
# create a stat_chart for the marker means
f.make_stat_chart(marker, 'marker_name', 'grade_mean', 'grade_mean')
# work through the readability stats to create a chart
for readability_list in cfg['analytics']['readability_stats']:
columns_old = ['marker_name']
columns_new = ['marker_name']
for i, row in comm.iterrows():
field = row['field']
text = row['label']
columns_old.append(field + "_" + readability_list[0])
columns_new.append(text)
this_marker = marker[columns_old].copy()
this_marker.columns = columns_new
f.make_stat_chart(this_marker, 'marker_name', columns_new, readability_list[0])
# start by creating a file to compile everything into
with open(c.d['md'] + cfg['analytics']['filename'] + ".md", 'w') as out:
print("## " + cfg['analytics']['analytics_header'] + "\n\n", file=out)
print("### " + cfg['analytics']['grade_table_header'] + "\n\n", file=out)
print("*" + cfg['analytics']['grade_table_comment'] + "*\n\n", file=out)
# create a summary table for display
overall_grade_count = c.df['marks']['grade_final'].replace(0, np.NaN).count().round(1)
overall_grade_mean = c.df['marks']['grade_final'].replace(0, np.NaN).mean().round(1)
overall_grade_min = c.df['marks']['grade_final'].replace(0, np.NaN).min().round(1)
overall_grade_max = c.df['marks']['grade_final'].replace(0, np.NaN).max().round(1)
overall_grade_std = c.df['marks']['grade_final'].replace(0, np.NaN).std().round(1)
overall_grade_skew = c.df['marks']['grade_final'].replace(0, np.NaN).skew().round(1)
marker_html = marker[
['marker_name', 'grade_count', 'grade_final', 'grade_std', 'grade_min', 'grade_max', 'grade_skew']].round(1)
overall_row = ["Overall", overall_grade_count, overall_grade_mean, overall_grade_std, overall_grade_min, overall_grade_max, overall_grade_skew]
marker_html.loc[-1] = overall_row
last_row = len(marker_html.index)
no_submission_count = len(c.df['marks'][c.df['marks']['grade_final'] == 0])
no_submission_row = ["No_Submission", no_submission_count, 'NA', 'NA', 'NA', 'NA', 'NA']
marker_html.loc[last_row] = no_submission_row # adding a row
marker_html = marker_html.sort_index() # sorting by index
marker_html.columns = ['Marker', 'Count', 'Mean', 'StDev', 'Min', 'Max', 'Skew']
marker_html.set_index('Marker', inplace=True)
print(marker_html.to_html(), file=out)
print("## Boxplot of final grades \n\n", file=out)
print("![](../../." + c.d['charts'] + "grade_final_pct_boxplot.png)\n\n", file=out)
print("## Histogram of final grades \n\n", file=out)
print("![](../../." + c.d['charts'] + "grade_final_pct.png)\n\n", file=out)
print("<div class=\"no-break\">\n\n", file=out)
print("## Difference between suggested and calculated grade \n\n", file=out)
print("*This highlights the tendancy of markers to drift from the suggested mark provided in the Database*\n\n", file=out)
print("![](../../." + c.d['charts'] + "count_suggested.png)\n\n", file=out)
print("</div>\n\n", file=out)
print("# " + cfg['analytics']['rubric_header'] + "\n\n", file=out)
for loop_row in crit.itertuples():
print("<div class=\"no-break\">\n\n", file=out)
criterion = loop_row.field
f.print_results_header(loop_row, out)
print('*' + cfg['analytics']['rubric_comment']+ '*\n\n', file=out)
print("![](../../." + c.d['charts'] + "count_" + criterion + ".png)\n\n", file=out)
print("</div>\n\n", file=out)
print("## " + cfg['analytics']['readability_header'] + "\n\n", file=out)
for readability_list in cfg['analytics']['readability_stats']:
print("<div class=\"no-break\">\n\n", file=out)
print("\n\n### " + cfg['crit_chart'][readability_list[0]], file=out)
print("![](../../." + c.d['charts'] + readability_list[0] + ".png)\n\n", file=out)
print("*" + cfg['crit_chart'][readability_list[0]] + cfg['analytics']['readability_comment'] + "*\n\n", file=out)
print("</div>\n\n", file=out)
print("\n\n## " + cfg['analytics']['sentiment_header'] + "\n\n", file=out)
sentiment_df = f.sentiment_table(comm, marker)
sentiment_df.set_index('Name', inplace=True)
print(sentiment_df.to_html(), file=out)
print("# " + cfg['analytics']['summary_header']+ "\n\n", file=out)
# combine the individual marker files
with open(c.d['md'] + cfg['analytics']['filename'] + '.md', 'a') as out_file:
for i, row in marker.iterrows():
this_marker_name = row['marker_name']
print(this_marker_name)
with open(c.d['md'] + this_marker_name + '.md') as in_file:
out_file.write(in_file.read())
with open(c.d['md'] + cfg['analytics']['filename'] + '.md', 'a') as out:
print("\n\n\n\n*** **END OF ANALYSIS** ***\n\n", file=out)
with open(c.d['yaml'] + cfg['analytics']['filename'] + '.yaml', 'w') as out:
f.pandoc_yaml(out, cfg['analytics']['filename'])
with open(c.d['css'] + cfg['analytics']['filename'] + ".css", 'w') as out:
f.pandoc_css(out, cfg['analytics']['filename'], 'anon')
f.pandoc_html_single(cfg['analytics']['filename'])
f.pandoc_pdf(cfg['analytics']['filename'])
f.pnt_notice(c.msg['console_complete'], os.path.basename(__file__))
