summaryrefslogtreecommitdiff
path: root/main.py
blob: ff63485e53d8c9fe0f3c23d4e394cc221ce6cdfe (plain)
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
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
import pandas as pd
import tkinter as tk
import textwrap
from tkinter import ttk
from tkinter.filedialog import askopenfilename, asksaveasfilename
from datetime import datetime

from ttkwidgets import ScrolledListbox

import sv_ttk
import darkdetect

import os
import json
import sys

import subprocess
import platform

from os import path

from docx import Document
from docx.shared import Inches
from docx.enum.text import WD_ALIGN_PARAGRAPH

from platformdirs import user_data_dir
from pathlib import Path

canadian_companies = []

# Config
appname = "PyramidScheme"
appauthor = "Navan Chauhan"
Path(user_data_dir(appname, appauthor)).mkdir( parents=True, exist_ok=True )

def get_companies_for_word_doc():
    None

def read_config(config_file='config.json'):
    default_config = {
        'img_width': 3300,
        'pyramid_width': 3000,
        'img_height': 1550,
        'overall_height': 2550,
        'block_color': [0, 150, 255],
        'dividend_color': [0, 150, 255],
        'font_color': [255, 255, 255],
        'normal_block': [0, 150, 255],
        'padding': 15,
        'radius': 10,
        'logo_padding': 100,
        'pyramid_padding_bottom': 350,
        'max_companies_in_row': 10,
    }
    
    if not os.path.exists(config_file):
        with open(config_file, 'w') as f:
            json.dump(default_config, f)
        return default_config
    else:
        with open(config_file, 'r') as f:
            config = json.load(f)
        return config

def pyramid_list(lst, sort_canadian=True):
    old_lst = lst

    top = []
    bottom = []

    for idx, company in enumerate(lst):
        if idx==0:
            top.append(company)
        else:
            if company in canadian_companies:
                bottom.append(company)
            else:
                top.append(company)

    lst = top + bottom

    i = 0
    rows = 0
    pyramid = []

    # Figure out the number of rows
    while i < len(lst):
        rows += 1
        i += rows

    # Generate the pyramid
    i = 0
    for r in range(1, rows+1):
        row = []
        for j in range(r):
            if i < len(lst):
                row.append(lst[i])
                i += 1
        pyramid.append(row)

    if len(pyramid) >= 2:
        if len(pyramid[-1]) < len(pyramid[-2]):
            pyramid[-2].extend(pyramid[-1])
            pyramid.pop()

    print("Checking for canadian companies")
    print(canadian_companies)
    if len(pyramid) > 2:
        print("Pyramid has more than two rows, checking to ensure Canadian companies are at the bottom")
        print(len(pyramid[-1]) - len(pyramid[-2]))
        for row in pyramid:
            print(len(row))
            print(row)
        if ((len(pyramid[-1]) - len(pyramid[-2])) > 1):
            print("Last two rows can be balanced if needed")
            for _ in range((len(pyramid[-1]) - len(pyramid[-2]))-1):
                company_to_add = None
                company_idx = 0
                for idx, company in enumerate(pyramid[-1]):
                    print(f'Checking {company}')
                    if company not in canadian_companies:
                        company_to_add = company
                        company_idx = idx
                        print(f'Shifting {company_to_add}')
                if company_to_add is not None:
                    pyramid[-1].pop(company_idx)
                    pyramid[-2].append(company_to_add)               

    for _ in range(4):
        if len(pyramid) > 2:
            if len(pyramid[-1]) < len(pyramid[-2]):
                pyramid[-3].append(pyramid[-2][-1])
                pyramid[-2].pop(-1)

    for _ in range(3):
        if len(pyramid) > 4:
            if len(pyramid[-2]) < len(pyramid[-3]):
                pyramid[-4].append(pyramid[-3][0])
                pyramid[-3].pop(0)

    return pyramid

class CompanySelector:
    def __init__(self, master):
        self.master = master
        self.master.title("Company Selector")
        self.master.geometry("600x600")
        self.select_file_button = ttk.Button(self.master, text="Select Excel File", command=self.load_file)
        self.select_file_button.pack(pady=250)
        print(path.join(user_data_dir(appname, appauthor), "default.txt"))
        # try:
        #     with open(path.join(user_data_dir(appname, appauthor), "default.txt")) as f:
        #         try:
        #             fname = f.read()
        #             self.df = pd.read_excel(fname)
        #             self.df = self.df[self.df['Company Name'].notna()]
        #             self.df = self.df[self.df['Symbol'].notna()]
        #             self.show_companies()
        #         except Exception as e:
        #             print(f'Oh No {e}')
        #             os.remove(path.join(user_data_dir(appname, appauthor), "default.txt"))
        #             self.load_file()
        # except FileNotFoundError:
        #     None
        self.show_companies()

    def load_file(self):
        file_path = askopenfilename(filetypes=[("Excel files", "*.xlsx *.xls")])
        if not file_path:
            return

        with open(path.join(user_data_dir(appname, appauthor), "default.txt"), "w") as f:
            f.write(file_path)
        try:
            self.df = pd.read_excel(file_path, keep_default_na=False)
            self.df = self.df[self.df['Company Name'].notna()]
            self.df = self.df[self.df['Symbol'].notna()]
        except Exception as e:
            print(f":( {e}")
            tk.messagebox.showerror(title="Whoops :(", message=f"{e}")
            return False
        return True
        #self.show_companies()


    def show_companies(self):
        try:
            with open(path.join(user_data_dir(appname, appauthor), "default.txt")) as f:
                try:
                    fname = f.read()
                    self.df = pd.read_excel(fname, keep_default_na=False)
                    self.df = self.df[self.df['Company Name'].notna()]
                    self.df = self.df[self.df['Symbol'].notna()]
                except Exception as e:
                    print(f'Oh No {e}')
                    tk.messagebox.showerror(title="Whoops! An Error Occurred", message=f"{e}\n\nYou will be asked to load the spreadsheet")
                    os.remove(path.join(user_data_dir(appname, appauthor), "default.txt"))
                    loaded_file = self.load_file()
                    while not loaded_file:
                        loaded_file = self.load_file()
        except FileNotFoundError:
            print("No default file set")
            tk.messagebox.showinfo(title="Hello world!", message="Since this is the first time you are using this software, please load the desired spreadsheet...")
            loaded_file = self.load_file()
            while not loaded_file:
                loaded_file = self.load_file()

        self.select_file_button.pack_forget()
        self.master.geometry("")
        self.master.resizable(width=False, height=False)

        width=600
        height=500
        screenwidth = self.master.winfo_screenwidth()
        screenheight = self.master.winfo_screenheight()
        alignstr = '%dx%d+%d+%d' % (width, height, (screenwidth - width) / 2, (screenheight - height) / 2)
        self.master.geometry(alignstr)
        self.master.title("UCS Pyramid Builder")


        self.company_listbox = ScrolledListbox(self.master, selectmode=tk.MULTIPLE, exportselection=False, height=20)
        self.company_listbox.place(x=10,y=20,width=235,height=457)
        #self.company_listbox.pack(expand=True)

        for ticker, company in zip(self.df["Company Name"],self.df["Symbol"]):
            self.company_listbox.listbox.insert(tk.END, f' {company} - {ticker}')


        # bind the listbox to an onselect event
        self.company_listbox.listbox.bind('<<ListboxSelect>>', self.onselect)

        self.num_companies_selected_label = ttk.Label(self.master, text="0", font='SunValleyBodyStrongFont 18 bold')
        self.num_companies_selected_label.place(x=315,y=370,width=30,height=30)

        self.companies_label = ttk.Label(self.master, text="Companies Selected", font='SunValleyBodyStrongFont 15')
        self.companies_label.place(x=340,y=370,width=200, height=30)

        self.pyramid_title_string = tk.StringVar()
        self.pyramid_title_string.set(f"Copyright (2004) by Gentry Capital Corporation")

        #self.pyramid_title =  ttk.Entry(self.master, textvariable=self.pyramid_title_string)
        #self.pyramid_title.pack()

        self.customization_label = ttk.Label(self.master, text="Customization", font='SunValleyBodyStrongFont 12 bold')
        self.customization_label.place(x=250, y=20, width=150, height=20)

        self.options_label = ttk.Label(self.master, text="Options", font='SunValleyBodyStrongFont 12 bold')
        self.options_label.place(x=250, y=150, width=100, height=20)

        fname = open(path.join(user_data_dir(appname, appauthor), "default.txt")).read()
        if len(fname) >= 35:
            fname = f"...{fname[-30:]}"
        self.db_label = ttk.Label(self.master, text=f"Using file: {fname}")
        self.db_label.place(x=250,y=250, width=329, height=40)

        self.prepared_for_string = tk.StringVar()
        self.prepared_for_entry = ttk.Entry(self.master, textvariable=self.prepared_for_string)
        self.prepared_for_entry.place(x=380,y=50,width=195,height=30)

        self.prepared_for_label = ttk.Label(self.master, text="Prepared For")
        self.prepared_for_label.place(x=250,y=50,width=132,height=30)

        self.advisor_string = tk.StringVar()
        self.advisor_entry = ttk.Entry(self.master, textvariable=self.advisor_string)
        self.advisor_entry.place(x=380,y=90,width=195,height=30)
        self.advisor_label = ttk.Label(self.master, text="Advisor")
        self.advisor_label.place(x=250,y=90,width=132,height=30)

        self.generate_pyramid_var = tk.IntVar()
        self.generate_pyramid_var.set(1)
        self.generate_pyramid_checkbx = ttk.Checkbutton(self.master, text="Generate Pyramid Image", onvalue=1, offvalue=0, variable=self.generate_pyramid_var)
        self.generate_pyramid_checkbx.place(x=250,y=180, width=329, height=30)

        self.generate_word_doc = tk.IntVar()
        self.generate_word_doc.set(0)
        self.generate_word_doc_checkbx = ttk.Checkbutton(self.master, text="Generate Description Document", onvalue=1, offvalue=0, variable=self.generate_word_doc)
        self.generate_word_doc_checkbx.place(x=250,y=220, width=329, height=30)

        self.load_another_file_button = ttk.Button(self.master, text="Load Different Spreadsheet", command=self.load_another_file)
        self.load_another_file_button.place(x=250, y=290, width=325, height=30)

        self.create_pyramid_button = ttk.Button(self.master, text="Create", command=self.create_pyramid)
        #self.create_pyramid_button.pack()
        self.create_pyramid_button.place(x=250,y=450,width=150,height=30)

        self.reset_button = ttk.Button(self.master, text="Reset", command=self.reset_fields)
        self.reset_button.place(x=250+150+25, y=450, width=150, height=30)

    def load_another_file(self):
        print("Trying to load another file...")
        loaded_file = self.load_file()
        self.show_companies()

    def onselect(self, evt):
        w = evt.widget
        num_selected = len(w.curselection())
        self.num_companies_selected_label.config(text=f"{num_selected}")

    def reset_fields(self):
        self.generate_word_doc.set(0)
        self.generate_pyramid_var.set(1)
        self.prepared_for_string.set("")
        self.advisor_string.set("")
        self.company_listbox.listbox.selection_clear(0,'end')
        self.num_companies_selected_label.config(text="0")

    def create_pyramid(self):
        selected_companies_tmp = [self.company_listbox.listbox.get(index) for index in self.company_listbox.listbox.curselection()]
        print(f'Selected {len(selected_companies_tmp)} companies')
        selected_companies = [x.split(" - ")[0].strip() for x in selected_companies_tmp]
        print(f'Symbols: {len(selected_companies)}')
        selected_df = self.df[self.df["Symbol"].isin(selected_companies)].sort_values(by="Weighting", ascending=False)
        print(f'DF: {len(selected_df)}')
        #pyramid_file_path = asksaveasfilename(defaultextension=".xlsx", filetypes=[("Excel files", "*.xlsx *.xls")])

        company_blurbs = {
        }

        if self.generate_word_doc.get() == 1:
            word_file_path = asksaveasfilename(defaultextension=".docx", filetypes=[("Word files", "*.docx")])

            if not word_file_path:
                return

            document = Document("assets/doc_template.docx")

            advisor = self.prepared_for_string.get()
            today = datetime.today()

            to_replace = document.paragraphs[1].text
            to_replace = to_replace.replace("{{ADVISOR}}", advisor)
            to_replace = to_replace.replace("{{DATE}}" , today.strftime("%B %m, %Y"))

            document.paragraphs[1].text = to_replace
            document.paragraphs[1].style.font.name = "Times New Roman"

            word_company_scores = {}

            for company, ticker, dividend, currency, desc, weighting in zip(selected_df["Company Name"], selected_df["Symbol"], selected_df["Dividend?"], selected_df["Currency"], selected_df["Blurb"], selected_df["Weighting"]):
                extra_char = ""
                if dividend.strip() == "Y":
                    extra_char = "***"
                if currency == "CAD":
                    canadian_companies.append(f'{company}{extra_char} ({ticker})')

                word_company_scores[f'{company}{extra_char} ({ticker})'] = weighting
                company_blurbs[f'{company}{extra_char} ({ticker})'] = desc

                # new_paragraph = document.add_paragraph()
                # temp_run = new_paragraph.add_run(f'{company}{extra_char} ({ticker})')
                # temp_run.bold = True
                # new_paragraph.alignment = WD_ALIGN_PARAGRAPH.LEFT
                # document.paragraphs.pop()
                # desc_paragraph = document.add_paragraph(f"\n{desc}\n")
                # desc_paragraph.alignment = WD_ALIGN_PARAGRAPH.LEFT
                # desc_paragraph.bold = False

            word_company_scores = {k: v for k, v in sorted(word_company_scores.items(), key=lambda item: item[1])}
            word_pyramid = {}
            for company, score in word_company_scores.items():
                if score not in word_pyramid:
                    word_pyramid[score] = []
                word_pyramid[score].append(company)

            sorted_keys = sorted(word_pyramid.keys())

            companies = []


            for key in sorted_keys:
                if word_pyramid[key] == []:
                    continue
                if len(word_pyramid[key])<=15:
                    companies.append(word_pyramid[key])
                else:
                    temp_list = []
                    for x in range(len(word_pyramid[key])//15):
                        temp_list.append(word_pyramid[key][x*15:15])

                    temp_list.append(word_pyramid[key][::-1][:len(word_pyramid[key])%15])
                    for row in temp_list[::-1]:
                        if row != []:
                            companies.append(row)

            everything_list = []
            for row in companies:
                if row == []:
                    continue
                for company in row:
                    everything_list.append(company)

            companies = everything_list

            companies = pyramid_list(companies)

            everything_list = []

            for row in companies:
                for company in row:
                    new_paragraph = document.add_paragraph()
                    temp_run = new_paragraph.add_run(company)
                    temp_run.bold = True
                    temp_run.font.name = "Times New Roman"
                    new_paragraph.alignment = WD_ALIGN_PARAGRAPH.LEFT
                    document.paragraphs.pop()
                    desc_paragraph = document.add_paragraph(f"\n{company_blurbs[company]}\n\n\n")
                    desc_paragraph.alignment = WD_ALIGN_PARAGRAPH.LEFT
                    desc_paragraph.bold = False
                    desc_paragraph.style.font.name = "Times New Roman"

            document.save(word_file_path)

        if self.generate_pyramid_var.get() == 0:
            None
        else: 
            image_file_path = asksaveasfilename(defaultextension=".png", filetypes=[("Image files","*.png")])

            company_scores = {}

            # For each 'Company Name', company_scores[company] = 'Weighting' for that company
            for company, ticker, weighting, dividend, currency in zip(selected_df["Company Name"], selected_df["Symbol"], selected_df["Weighting"], selected_df["Dividend?"], selected_df["Currency"]):
                extra_char = ""
                if dividend.strip() == "Y":
                    extra_char = "*"
                if currency == "CAD":
                    canadian_companies.append(f'{company} ({ticker}){extra_char}')
                company_scores[f'{company} ({ticker}){extra_char}'] = weighting

            from PIL import Image, ImageDraw, ImageFont
            import math
            # Sort the dictionary by score in ascending order
            company_scores = {k: v for k, v in sorted(company_scores.items(), key=lambda item: item[1])}

            # Group the companies by their scores
            pyramid = {i: [] for i in range(1, 7)}
            for company, score in company_scores.items():
                pyramid[score].append(company)

            # Initialize some parameters
            img_width = 3300
            pyramid_width = 3000
            img_height = 1550 #2550
            overall_height = 2550
            block_color = (0, 150, 255)  # Blue color
            dividend_color = block_color #(1,50,32)  # Green color
            font_color = (255, 255, 255)  # White color
            normal_block = (0, 150, 255) # Dark Blue  
            padding = 15  # Padding around blocks
            radius = 10 
            logo_padding = 100 # top padding for logo
            pyramid_padding_bottom = 350 # bottom padding for pyramid
            max_companies_in_row = 10

            config = read_config(path.join(user_data_dir(appname, appauthor), "config.json"))
            print(config)
            img_width = config['img_width']
            pyramid_width = config['pyramid_width']
            img_height = config['img_height']
            overall_height = config['overall_height']
            block_color = tuple(config['block_color'])
            dividend_color = tuple(config['dividend_color'])
            font_color = tuple(config['font_color'])
            normal_block = tuple(config['normal_block'])
            padding = config['padding']
            radius = config['radius']
            logo_padding = config['logo_padding']
            pyramid_padding_bottom = config['pyramid_padding_bottom']
            max_companies_in_row = config['max_companies_in_row']

            base_path = "assets"
            primary_font = path.join(base_path, "baskerville.ttf")
            secondary_font = path.join(base_path, "gill_sans_bold.ttf")

            # Group the companies by their scores
            pyramid = {}
            for company, score in company_scores.items():
                if score not in pyramid:
                    pyramid[score] = []
                pyramid[score].append(company)

            # Sort the pyramid keys in ascending order to have a pyramid shape
            sorted_keys = sorted(pyramid.keys())

            companies = []


            for key in sorted_keys:
                if pyramid[key] == []:
                    continue
                if len(pyramid[key])<=max_companies_in_row:
                    companies.append(pyramid[key])
                else:
                    temp_list = []
                    for x in range(len(pyramid[key])//max_companies_in_row):
                        temp_list.append(pyramid[key][x*max_companies_in_row:max_companies_in_row])

                    temp_list.append(pyramid[key][::-1][:len(pyramid[key])%max_companies_in_row])
                    for row in temp_list[::-1]:
                        if row != []:
                            companies.append(row)

            everything_list = []
            for row in companies:
                if row == []:
                    continue
                for company in row:
                    everything_list.append(company)

            companies = everything_list
            print(f"Still have {len(companies)}")

            companies = pyramid_list(companies)

            for company_row in companies:
                print(company_row)

            # Calculate the maximum number of companies in a group (this will be the width of your pyramid)
            max_companies = max(len(row) for row in companies) #max(len(v) for v in pyramid.values())

            # Calculate the total number of groups (the height of your pyramid)
            total_groups = len(companies) #len(pyramid)

            # Calculate the size of each block based on the width and height of the image and the number of blocks
            block_size = min((pyramid_width - padding) // max_companies - padding, (img_height - padding) // total_groups - padding)

            # Calculate the size of each block based on the width and height of the image and the number of blocks
            block_width = (pyramid_width - padding) // max_companies - padding
            block_height = (img_height - padding) // total_groups - padding

            # Calculate the total width and height of the blocks (including padding)
            total_width = max_companies * (block_width + padding)
            total_height = total_groups * (block_height + padding)

            # Calculate the starting position for the first block
            start_x = (pyramid_width - total_width) // 2
            start_y = (img_height - total_height) // 2


            # Create an image big enough to hold the pyramid
            img = Image.new('RGB', (pyramid_width, img_height), "white")
            d = ImageDraw.Draw(img)

            def wrap_text(text, max_length):
                words = text.split()
                lines = []
                current_line = []

                for word in words:
                    if len(' '.join(current_line + [word])) <= max_length:
                        current_line.append(word)
                    else:
                        lines.append(' '.join(current_line))
                        current_line = [word]
                lines.append(' '.join(current_line))
                
                return '\n'.join(lines)

            min_font_size = 100_000_000

            for i, row in enumerate(companies):
                for j, company in enumerate(row):
                    font_size = min(block_width // (len(company) // 2 + 1), block_height // 2)
                    if font_size < min_font_size:
                        min_font_size = font_size

            num_dividends = 0

            # Loop over each level of the pyramid
            for i, row in enumerate(companies):
                for j, company in enumerate(row):
                    # Calculate the position of the block
                    x = start_x + j * (block_width + padding) + (max_companies - len(row)) * (block_width + padding) // 2
                    y = start_y + i * (block_size + padding)

                    # Calculate the color of the block
                    if company[-1] == "*":
                        block_color = dividend_color
                        num_dividends += 1
                    else:
                        block_color = normal_block

                    # Draw the block
                    d.rounded_rectangle([x, y, x + block_width, y + block_height], fill=block_color, radius=radius)

                    # Adjust font size based on the length of the company name and block size
                    font_size = min_font_size + 15
                    fnt = ImageFont.truetype(secondary_font, font_size)

                    # Implement word wrap for the company name
                    wrapped_company = wrap_text(company, (block_width // font_size)*1.5)

                    # Draw the company name
                    bbox = d.textbbox((x, y), wrapped_company, font=fnt)
                    text_width = bbox[2] - bbox[0]
                    text_height = bbox[3] - bbox[1]
                    text_x = x + (block_width - text_width) // 2
                    text_y = y + (block_height - text_height) // 2
                    if "Constellation" in company and len(row) == 5:
                        print("Manually shifting up")
                        text_y -= 25
                    d.text((text_x, text_y), wrapped_company, font=fnt, fill=font_color, align='center', spacing=10)


            print(image_file_path)

            # Save the image
            if not image_file_path:
                return

            image_logo = Image.open("./assets/GentryCapitalRGB.jpg")
            _, l_height = image_logo.size

            new_img = Image.new("RGB", (img_width, l_height), "white")
            left = (new_img.width - image_logo.width) // 2
            top = (new_img.height - image_logo.height) // 2

            new_img.paste(image_logo, (left, top))

            final_img = Image.new("RGB", (img_width, overall_height), "white")
            final_img.paste(new_img, (0, logo_padding))

            text1 = self.pyramid_title_string.get()
            text2 = f"({num_dividends} Dividend payors - all identified by asterisk)"
            text3 = "FOR INTERNAL USE ONLY"
            text4 = datetime.today().strftime("%B %d, %Y")
            #text5 = f"{len([self.company_listbox.listbox.get(index) for index in self.company_listbox.listbox.curselection()])}"
            text5 = f"Number of positions: {len([self.company_listbox.listbox.get(index) for index in self.company_listbox.listbox.curselection()])}"

            if self.prepared_for_string.get() != "":
                print(f"Prepared for string present - {self.prepared_for_string.get()}")
                if text5 != "":
                    if type(text5) == list:
                        text5.append(f"Prepared for {self.prepared_for_string.get()}")
                    else:
                        temp_prepared_for_text = f"Prepared for {self.prepared_for_string.get()}"
                        if len(temp_prepared_for_text) > 30:
                            text5 = [text5, "Prepared for: "]
                            for line in textwrap.wrap(self.prepared_for_string.get(), 30, break_long_words=False):
                                text5.append(line)
                        else:
                            text5 = [text5, f"Prepared for {self.prepared_for_string.get()}"]
            if self.advisor_string.get() != "":
                advisor = self.advisor_string.get()
                if text5 != "":
                    if type(text5) == list:
                        text5.append(f"By {advisor}")
                    else:
                        tmp_var = [text5, f"By {advisor}"]
                        text5 = tmp_var
                else:
                    text5 = f"By {advisor}"

            print(text5)
            draw = ImageDraw.Draw(final_img)
            font_size = 60

            font1 = ImageFont.truetype(primary_font, font_size)
            text_width1 = draw.textlength(text1, font=font1)
            while text_width1 > final_img.width:
                font_size -= 1
                font1 = ImageFont.truetype(primary_font, font_size)
                text_width1 = draw.textlength(text1, font=font1)

            font2_size = font_size - 10
            font2 = ImageFont.truetype(primary_font, font2_size)
            text_width2 = draw.textlength(text2, font=font2)
            while text_width2 > final_img.width:
                font_size -= 1
                font2 = ImageFont.truetype(primary_font, font2_size)
                text_width2 = draw.textlength(text2, font=font2)

            new_font_size = 30
            font3 = ImageFont.truetype(secondary_font, new_font_size)
            _,t,_,b = draw.textbbox((100,100), text3, font=font3)
            while (b-t) < (0.4*pyramid_padding_bottom):
                new_font_size += 1
                font3 = ImageFont.truetype(secondary_font, new_font_size)
                _,t,_,b = draw.textbbox((100,100), text3, font=font3)
            text_width3 = draw.textlength(text3, font=font3)
            while text_width3 > (final_img.width*0.8):
                new_font_size -= 1
                font3 = ImageFont.truetype(secondary_font, new_font_size)
                text_width3 = draw.textlength(text3, font=font3)
            _,t,_,b = draw.textbbox((100,100),text3,font=font3)

            font4_size = font2_size - 5
            font4 = ImageFont.truetype(primary_font, font4_size)
            text_width4 = draw.textlength(text4, font=font4)


            start_x1 = (final_img.width - text_width1) // 2
            start_y1 = image_logo.height + logo_padding + 50

            start_x2 = (final_img.width - text_width2) // 2
            start_y2 = image_logo.height + logo_padding + 125

            start_x3 = (final_img.width - text_width3) // 2
            start_y3 = (final_img.height - pyramid_padding_bottom + ((pyramid_padding_bottom)-(b-t))//2)

            start_x4 = (final_img.width - text_width4) - 150
            start_y4 = logo_padding + 30 

            start_x5 = 150
            start_y5 = start_y4
            
            draw.text((start_x1, start_y1), text1, font=font1, fill="black")
            draw.text((start_x2, start_y2), text2, font=font2, fill="black")
            draw.text((start_x3, start_y3), text3, font=font3, fill=(158,161,162))
            draw.text((start_x4, start_y4), text4, font=font4, fill="black")
            if type(text5) == str:
                draw.text((start_x5, start_y5), text5, font=font4, fill="black")
            else:
                for idx, text2write in enumerate(text5):
                    draw.text((start_x5, start_y5 + (idx*69)), text2write, font=font4, fill="black")

            final_img.paste(img, ((img_width-pyramid_width)//2, l_height + (overall_height - l_height - img_height) - pyramid_padding_bottom) )

            final_img.save(image_file_path)
            #final_img.show()
            #img.show()

            if platform.system() == 'Darwin':       # macOS
                subprocess.call(('open', image_file_path))
            elif platform.system() == 'Windows':    # Windows
                os.startfile(image_file_path)
            else:                                   # linux variants
                subprocess.call(('xdg-open', image_file_path))


if __name__ == "__main__":
    root = tk.Tk()
    wd = os.getcwd()
    try:
        os.chdir(sys._MEIPASS)
    except AttributeError:
        os.chdir(wd)
    print(os.listdir())
    app = CompanySelector(root)
    sv_ttk.set_theme("dark")
    if darkdetect.isLight():
        sv_ttk.set_theme("light")

    root.mainloop()