def plot_totals(ts_dict):
"""
Plot the total statistics (number of words, number of texts) per person.
Args:
ts_dict (dict): {"PersonName": TextSummary for that person}.
Returns:
None. Opens plot.
"""
total_attrs = ['texts', 'words']
titles = ['Texts', 'Words']
number_of_plots = len(total_attrs)
fig = make_subplots(rows=number_of_plots,
cols=1,
shared_xaxes=False,
subplot_titles=(titles))
for index in range(number_of_plots):
for sender in list(COLORS.keys()):
fig.append_trace(
go.Bar(x=[titles[index]],
y=[ts_dict[sender].count[total_attrs[index]]],
name=sender,
showlegend=True,
marker_color=COLORS[sender]), index + 1, 1)
fig.update_layout(width=600, height=400 * number_of_plots, barmode='group')
fig.update_yaxes(title_text='Total')
plot(fig, auto_open=True, filename="Total Stats.html")
def sub_parser_add(sub_parser):
add_parser = sub_parser.add_parser('add', help="Add event to calendar")
add_parser.add_argument('-c', '--calendar', type=str, help="Calendar name")
add_parser.add_argument('title', type=str, help="Event title")
add_parser.add_argument('start_date',
type=valid_date,
help="Event start time, format: YYYY-MM-DD")
add_parser.add_argument('start_time',
type=valid_time,
help="Event start time, format: HH:MM")
add_parser.add_argument('-d',
'--duration',
type=int,
help="Event duration (minutes)")
add_parser.add_argument('-a',
'--attendees',
type=valid_attendees,
help="List of emails of attendees",
required=False)
add_parser.add_argument('-o',
'--override-color',
type=str,
choices=[c for c in sorted(COLORS.keys())],
help="List of emails of attendees",
default="")
def prepare_summary(self, n_keywords: int = 4):
"""
Now prepare data for summary page
The output is [(topic_idx, [(word, weight, color_in_hex), ...]), ...]
"""
# `temp_all_topic_word_weight_sent` = [(topic_idx, word, weight, sentiment), ...]
temp_all_topic_word_weight_sent = []
for topic in self.topic_words:
temp_list = []
for ww in topic[1]:
if ww[0] in self.sent.keys():
temp_list.append(
(topic[0], ww[0], ww[1], self.sent[ww[0]]))
temp_all_topic_word_weight_sent.extend(temp_list)
# sort by sentiment
temp_all_topic_word_weight_sent.sort(key=lambda x: x[3], reverse=True)
# initialize topic_summary [(topic_idx, topic_sent_value, topic_keywords, [(word,weight,color_hex), ...]), ...]
topic_summary = []
for idx in range(self.n_topics):
topic_summary.append([idx, None, None, []])
for x in enumerate(temp_all_topic_word_weight_sent):
color = list(COLORS.keys())[int(
float(x[0]) / len(temp_all_topic_word_weight_sent) * 7.999)]
topic_summary[x[1][0]][3].append((
x[1][1], # word
np.float64(x[1][2]), # weight
color, # sentiment color
))
def get_topic_sent(t):
d = {}
for w in t:
if w[2] in d.keys():
d[w[2]] += w[1]
else:
d[w[2]] = w[1]
return sorted(d.items(), key=lambda x: x[1], reverse=True)[0][0]
for topic in topic_summary:
# topic_sent_value color string
topic[1] = get_topic_sent(topic[3])
# topic_keywords
topic[2] = [
kw[0] for kw in sorted(
topic[3], key=lambda x: x[1], reverse=True)[:n_keywords]
]
# Save the data
self._dump_intermediate(f"{self.topic_summary_fname}.pkl",
topic_summary)
x = [c['duration'].delta / 60e9 for c in convos]
y = [c['total_words'] for c in convos]
fig = go.Figure(data=go.Scatter(x=x, y=y, mode='markers'))
fig.update_layout(title='Convo duration vs words',
xaxis_title="Duration (minutes)",
yaxis_title="Words",
yaxis_scaleanchor="x",
yaxis_scaleratio=1)
fig.update_xaxes(type='log')
fig.update_yaxes(type='log')
plot(fig, auto_open=True, filename="Convo Duration vs Words.html")
if __name__ == "__main__":
START_TIME = time.time()
df = pd.read_csv(LOADPATH, index_col=False)
df = convert_raw_csv_types(df)
names = list(COLORS.keys())
df_person_list = [df[df['sender'] == name] for name in names]
summary_all = TextSummary(df)
ts_dict = dict(
zip(names, [TextSummary(df_person) for df_person in df_person_list]))
plot_totals(ts_dict)
conversations = summary_all.get_conversations(names)
plot_convo_words(conversations, names)
plot_convo_length(conversations)
print("--- %s sec execution time ---" % (time.time() - START_TIME))
def make_scatter_plots(text_summary_by_person):
"""
Create scatter plot of ratio of each person's usage of words
Args:
text_summary_by_person (dict): {"Person1": TextSummary for that person, etc.}.
Returns:
None. Opens plot.
"""
MIN_WORD_OCCUR = 10 # minimum number of occurrences to be plotted
MIN_EMOTE_OCCUR = 5
names = list(COLORS.keys())
diffs = dict() # stores dictionaries with key word and value (total, ratio)
# make dictionaries
diffs['words'] = text_summary_by_person[0].compare_freq(text_summary_by_person[1], 'words')
diffs['emotes'] = text_summary_by_person[0].compare_freq(text_summary_by_person[1], 'emotes')
# strip less frequent tokens
diffs['words'] = {key: value for (key, value) in diffs['words'].items() if
value[0] >= MIN_WORD_OCCUR}
diffs['emotes'] = {key: value for (key, value) in diffs['emotes'].items() if
value[0] >= MIN_EMOTE_OCCUR}
number_of_plots = 2
titles = ['Words', 'Emotes']
# initialize subplot figure
fig = make_subplots(rows=number_of_plots, cols=1, shared_xaxes=True,
subplot_titles=(titles))
subplot = 0
for token_type, diff_dict in diffs.items():
x, y, customdata = [], [], []
for word, v in diff_dict.items():
y.append(v[0])
x.append(v[1])
customdata.append([word, names[0], v[2], names[1], v[3]])
fig.append_trace(go.Scatter(x=x,
y=y,
customdata=customdata,
hovertemplate="<b>%{customdata[0]}</b><br>" +\
"Total: %{y} <br>Ratio: %{x:.2f}<br>" +\
"%{customdata[1]}: %{customdata[2]}<br>" +\
"%{customdata[3]}: %{customdata[4]}<extra></extra>",
showlegend=False,
mode='markers'),
subplot+1, 1)
subplot += 1
fig.update_layout(width=1200, height=600*number_of_plots)
fig.update_xaxes(type='log', title_text=f'{names[0]}: {names[1]} ratio')
fig.update_yaxes(type='log', title_text='Total')
plot(fig, auto_open=True, filename="Word Ratio Scatterplot.html")