mapp_streamlit/main.py at main · zarinaIbr/mapp_streamlit · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
from collections import defaultdict
import base64
import pandas as pd
import pickle
import plotly.graph_objs as go
import pygsheets
import streamlit as st


st.markdown(
    """<style>
    [data-testid="stSidebar"][aria-expanded="true"] > div:first-child {width: 370px;}
    [data-testid="stSidebar"][aria-expanded="false"] > div:first-child {width: 1000px; margin-left: -500px;}
    </style>""", unsafe_allow_html=True)

st.markdown("<h2 style='text-align: center; color: black;'>Генерация правил для исправления ААО</h2>",
            unsafe_allow_html=True)

@st.cache(allow_output_mutation=True)
def load_data():
    with open('data_mapp_web_m50.pickle', 'rb') as f:
        data = pickle.load(f)
        return data, data.loc[:, ['query', 'freq']]

def svg_html(reac):
    svg = reac.depict()
    b64 = base64.b64encode(svg.encode('utf-8')).decode("utf-8")
    html = r'<img src="data:image/svg+xml;base64,%s"/>' % b64
    return html

def update_reaction(old_r, f):
    new_r = old_r.copy()
    for mol in new_r.products:
        mol.remap(f)
    return new_r

def gen_fix_r(reaction, num_bad, num_good):
    fix = {t[0]: t[1] for t in zip(num_bad, num_good)}
    reaction_new = update_reaction(reaction, fix)
    return fix, reaction_new

def load_remapping_rules(reactions):
    rules = []
    for bad, good in reactions:
        if str(bad) != str(good):
            raise ValueError('bad and good reaction should be equal')

        cgr_good, cgr_bad = ~good, ~bad
        gc = cgr_good.augmented_substructure(cgr_good.center_atoms, deep=1)
        bc = cgr_bad.augmented_substructure(cgr_bad.center_atoms, deep=1)
        atoms = set(bc.atoms_numbers + gc.atoms_numbers)

        re_g, re_b, pr_g, pr_b = set(), set(), set(), set()
        for pr in good.reactants:
            re_g.update(pr)
        for pr in good.products:
            pr_g.update(pr)
        for pr in bad.products:
            pr_b.update(pr)
        for pr in bad.reactants:
            re_b.update(pr)
        atoms.update((re_b.difference(pr_b)).intersection(pr_g))
        strange_atoms = pr_b.difference(pr_g)
        atoms.update(strange_atoms)
        bad_query = (cgr_bad).substructure(atoms.intersection(cgr_bad), as_query=True)
        good_query = (cgr_good).substructure(atoms.intersection(cgr_good), as_query=True)

        fix = {}
        for mb, mg in zip(bad.products, good.products):
            fix.update({k: v for k, v in zip(mb, mg) if k != v and k in atoms})  # get fix map
        valid = set(fix).difference(strange_atoms)
        rules.append((bad_query, good_query, fix, valid))
    return rules

def download_to_excel(d_info, type_aam):
    path = 'main_secret.json'
    gc = pygsheets.authorize(service_file=path)
    if type_aam:
        sh = gc.open('good_mapping')
    else:
        sh = gc.open('bad_mapping')
    wks = sh.worksheet_by_title('Лист1')
    df_excel = wks.get_as_df()
    df_new = pd.DataFrame(d_info)
    values = df_new[['rc', 'index', 'freq']].values.tolist()
    wks.append_table(values, start=f'A{len(df_excel) + 1}', end=None, dimension='ROWS', overwrite=False)

data_all, data_df = load_data()
st.sidebar.write('Датасет', data_df)
selected_indices = st.sidebar.multiselect('Выберите р.ц', data_df.index)
selected_rows = data_df.loc[selected_indices]
st.sidebar.write('Выбранный р.ц:', selected_rows)

st.sidebar.markdown("<h2 style='text-align: center; color: black;'>Инструкция</h2>", unsafe_allow_html=True)
st.sidebar.markdown("""
* 1.Выберите один р.ц по его индексу в таблицe и проверьте его правильность по реакции
* 2.Если маппинг правильный, нажмите Да, и переходите к следующему р.ц
* 3.Если маппинг ошибочный, нажмите Нет
    * 3.1.Если словарь НЕ пустой, очистите словарь, нажмите Delete!!!
    * 3.2.Добавляйте пары атомов в пустой словарь, нажимая на Add
* 4.После сбора словаря, проверьте себя по сгенерированной реакции
* 5.Если исправленный маппинг реакции вы считаете верным, нажмите Сгенерировать правило
* 6.После окончания сбора правил, нажмите "Save true/false data" и "Сохранить данные"
""")

if len(selected_indices) == 1:
    reaction = data_all['reaction'].values[int(selected_indices[0])][0]
    st.markdown("<p style='text-align: center; color: black;'>Исходная реакция</p>", unsafe_allow_html=True)
    st.write(svg_html(reaction), unsafe_allow_html=True)

    st.markdown("<p style='text-align: left; color: black;'>ААО верное?</p>", unsafe_allow_html=True)
    checkbox_yes = st.checkbox("Да, корректное")
    checkbox_no = st.checkbox("Нет, ошибочное")
    if checkbox_no and checkbox_yes:
        st.error('Выберите только один вариант')
    elif checkbox_yes and not checkbox_no:
        if 'good_info' not in st.session_state:
            st.session_state.good_info = defaultdict(list)
        if selected_rows.values.tolist()[0][0] not in st.session_state.good_info['rc']:
            st.session_state.good_info['index'].append(int(selected_indices[0]))
            st.session_state.good_info['rc'].append(selected_rows.values.tolist()[0][0])
            st.session_state.good_info['freq'].append(selected_rows.values.tolist()[0][1])
            st.session_state.good_info['nums_reac'].append(data_all['nums_reactions'].values[int(selected_indices[0])])
    elif checkbox_no and not checkbox_yes:
        if 'before' not in st.session_state and 'after' not in st.session_state:
            st.session_state.before, st.session_state.after = [], []
        col1, col2 = st.columns(2)
        with col1:
            sym_b = st.number_input("Ошибочный номер атома", min_value=1)
            sym_a = st.number_input("Корректный номер атома", min_value=1)
        with col2:
            st.write('Сохранить пару')
            add_button = st.button('Add')
            st.write('Очистить словарь')
            del_button = st.button('Delete')
        if add_button:
            st.session_state.before.append(sym_b)
            st.session_state.after.append(sym_a)
        elif del_button:
            while len(st.session_state.after) > 0:
                st.session_state.after.pop()
                st.session_state.before.pop()
        st.code({t[0]: t[1] for t in zip(st.session_state.before, st.session_state.after)})
        col_r, col_rule = st.columns(2)
        with col_r:
            show_reac = st.button('Показать реакцию после исправления ААО')
        with col_rule:
            load_rule = st.button('Сгенерировать правило')
        if show_reac:
            try:
                fix_web, reaction_new = gen_fix_r(reaction, st.session_state.before, st.session_state.after)
                st.write(svg_html(reaction_new), unsafe_allow_html=True)
            except:
                st.error('Ошибка при исправлении ААО')
        if load_rule:
            fix_web, reaction_new = gen_fix_r(reaction, st.session_state.before, st.session_state.after)
            if 'bad_info' not in st.session_state:
                st.session_state.bad_info = defaultdict(list)
            if selected_rows.values.tolist()[0][0] not in st.session_state.bad_info['rc']:
                try:
                    rule = load_remapping_rules([(reaction, reaction_new)])
                    st.session_state.bad_info['index'].append(int(selected_indices[0]))
                    st.session_state.bad_info['rc'].append(selected_rows.values.tolist()[0][0])
                    st.session_state.bad_info['freq'].append(selected_rows.values.tolist()[0][1])
                    st.session_state.bad_info['nums_reac'].append(data_all['nums_reactions'].values[int(selected_indices[0])])
                    st.session_state.bad_info['bad_r'].append(reaction)
                    st.session_state.bad_info['good_r'].append(reaction_new)
                    st.session_state.bad_info['rule'].append(rule)

                    col_bq, col_gq, col_f = st.columns([5, 5, 1])
                    with col_bq:
                        st.text('Р.ц непр.ААО')
                        st.write(svg_html(rule[0][0]), unsafe_allow_html=True)
                    with col_gq:
                        st.text('Р.ц прав.ААО')
                        st.write(svg_html(rule[0][1]), unsafe_allow_html=True)
                    with col_f:
                        st.text('Словарь пар атомов')
                        st.write(f'{rule[0][2]}')
                except:
                    st.error('Ошибка при создании правила.')

stat = st.button('Посмотреть статистику')
if stat:
    with st.expander("Свернуть/Развернуть", True):
        path = 'main_secret.json'
        gc = pygsheets.authorize(service_file=path)
        sh_good = gc.open('good_mapping')
        sh_bad = gc.open('bad_mapping')
        wks_good = sh_good.worksheet_by_title('Лист1')
        wks_bad = sh_bad.worksheet_by_title('Лист1')
        df_bad, df_good = wks_bad.get_as_df(), wks_good.get_as_df()

        col1, col2, col3 = st.columns(3)
        col1.metric("Всего реакций", "1358000")
        col2.metric("Кол-во правильных реакций", f"{df_good['freq'].sum()}")
        col3.metric("Создано руллов", f"{len(df_bad['freq'])}")

        labels = ['Корректный маппинг', 'Потенциально корректный маппинг', 'Неизвестно']
        values = [df_good['freq'].sum(), df_bad['freq'].sum(), 1358000 - df_good['freq'].sum() - df_bad['freq'].sum()]
        colors = ['green', 'yellow', 'gray']
        fig = go.Figure(data=[go.Pie(labels=labels, values=values)])
        fig.update_traces(marker=dict(colors=colors))
        st.plotly_chart(fig, use_container_width=True)

but_t = st.sidebar.button('Save True data')
but_f = st.sidebar.button('Save False data')
if but_t:
    out = pickle.dumps(st.session_state.good_info)
#     data_good = pd.DataFrame(st.session_state.good_info)
#     csv_good = data_good.to_csv().encode('utf-8')
    download_to_excel(st.session_state.good_info, type_aam=True)
    st.sidebar.download_button(label="Сохранить данные по правильным реакциям", data=out, file_name=f"good_data_{int(selected_indices[0])}.pickle")
    del st.session_state.good_info
if but_f:
    out_bad = pickle.dumps(st.session_state.bad_info)
    download_to_excel(st.session_state.bad_info, type_aam=False)
    st.sidebar.download_button(label="Скачать данные по ошибочным реакциям", data=out_bad, file_name=f"bad_data_{int(selected_indices[0])}.pickle")
    del st.session_state.bad_info