File size: 23,586 Bytes
b073d38 41ace82 b073d38 5199ad8 b073d38 |
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 |
from reportlab.lib.pagesizes import A4
from reportlab.lib import colors
from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle
from reportlab.lib.units import inch, cm
from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, Table, TableStyle
from reportlab.lib.enums import TA_LEFT, TA_RIGHT, TA_CENTER
from reportlab.pdfgen import canvas
from reportlab.platypus import Image
import streamlit as st
import pandas as pd
from io import BytesIO
import xlsxwriter
from datetime import date, datetime
from st_aggrid import AgGrid
from st_aggrid.grid_options_builder import GridOptionsBuilder
from helpertools_TM import helpers
class ReportGenerator:
def __init__(self, processor, trova_errori, interface, no_zero=None):
"""Initialize ReportGenerator with required components.
processor: Handles data processing and error collection
trova_errori: Contains error detection logic and dataframes
interface: Handles UI interactions and grid displays
no_zero: Flag for handling zero values
"""
self.df = trova_errori.df
self.df_bambini = trova_errori.df_bambini_che_impediscono_controllo_superamento_ore
self.df_calcolo_666 = trova_errori.df_calcolo_666
self.dffinal = trova_errori.dffinal
self.processor = processor
self.no_zero = no_zero
self.interface = interface
self.trova_errori = trova_errori
self.ERRORDICT = trova_errori.ERRORDICT
self.helper = helpers(interface,self.ERRORDICT)
def get_version(self):
return "ReportGenerator v1.0.1"
@st.fragment()
def display_final_table(self) -> None:
"""
Displays the final table with all data and a separate table for errors only.
This method creates two expanders: one for displaying all data in a grid format,
and another for showing only records that contain errors. It also provides options
to download these tables as Excel files.
Args:
None
Returns:
None
"""
# Display the final table with all data
no_zero = self.no_zero # Assuming this is defined somewhere else in the code.
# Create expander for full data table
# This section shows all records, including those without errors
expndr = st.expander("TABELLA FINALE ELABORATA - TUTTI I DATI")
with expndr:
# Enable Excel download of complete dataset
# Sort by surname and name for better readability
self.interface.download_excel_file(
self.df.sort_values(by=["Cognome", "Nome"]),
"TabellaFinaleElaborata_TuttiDati.xlsx",
'mantieniColonneErrore'
)
# Configure and display interactive grid with all records
# Enable enterprise features for better user interaction
gridOptions = self.interface.buildGrid(self.df.sort_values(by=["Cognome", "Nome"]))
AgGrid(
self.df.sort_values(by=["Cognome", "Nome"]),
gridOptions=gridOptions,
enable_enterprise_modules=True,
)
# Display the table with only errors
# Create expander for error-only table
# This section filters and shows only records with detected errors
expndr = st.expander("TABELLA FINALE ELABORATA - SOLO ERRORI")
with expndr:
# Filter data to show only records with errors
self.dffinal = self.helper.make_df_solo_errori(self.df)
# Provide download option for the errors-only table
self.interface.download_excel_file(
self.dffinal.sort_values(by=["Cognome", "Nome"]),
"TabellaFinaleElaborata_SoloErrori.xlsx",
"mantieniColonneErrore"
)
# Display the errors-only table in a grid format
gridOptions = self.interface.buildGrid(self.dffinal.sort_values(by=["Cognome", "Nome"]))
AgGrid(
self.dffinal.sort_values(by=["Cognome", "Nome"]),
gridOptions=gridOptions,
enable_enterprise_modules=True,
key="tabfinalex"
)
# Save the errors-only table to an Excel file
# try:
# self.dffinal.to_excel("storico.xlsx")
# st.success("Salvato file storico: OK")
# except Exception as e:
# st.error(f"ERRORE salvataggio file storico: errore({e})")
def genera_report(self) -> None:
"""
Generates detailed reports in Excel and PDF formats and provides download options via Streamlit.
This method filters errors, computes unique children and municipalities, and generates
Excel and PDF reports if errors are present. It handles various exceptions that may occur
during the report generation process.
Args:
None
Returns:
None
"""
# Generate detailed reports in Excel and PDF formats
try:
# Filter out specific errors and get current error state
error_keys, errors_present = self._filter_errors(excluded_error="errMassimo543")
# Early return if no checks are needed
if not error_keys:
st.warning("Nessun controllo da eseguire.")
return
# Calculate statistics about the dataset
nr_bambini = self._compute_unique_children()
nr_comuni = self._compute_unique_gemeinden()
# Early return if no errors were found
if not errors_present:
st.info("Nessun errore rilevato nei controlli selezionati.")
return
# Generate reports and provide download options
try:
excel_data = self._generate_excel_report(nr_bambini, nr_comuni, errors_present)
except (xlsxwriter.exceptions.FileCreateError, PermissionError) as e:
st.error(f"File access error: {str(e)}")
st.stop()
except ValueError as ve:
st.error(f"Invalid data format: {str(ve)}")
st.stop()
except KeyError as ke:
st.error(f"Missing required column: {str(ke)}")
st.stop()
self._provide_downloads(excel_data)
st.success("Generato file riassunto analisi")
except KeyError as ke:
# Handle missing or invalid data structure errors
st.error(f"Chiave mancante o struttura dati non valida: {ke}")
st.stop()
except ValueError as ve:
st.error(f"Problema di tipo o formato dati: {ve}")
st.stop()
except FileNotFoundError as fnf:
st.error(f"File di riferimento non trovato: {fnf}")
except PermissionError as pe:
st.error(f"Permessi insufficienti per accedere al file: {pe}")
except Exception as e:
st.error(f"Errore imprevisto durante la generazione del report: {e}")
st.stop()
def _filter_errors(self, excluded_error: str):
"""
Excludes a specific error from the error dictionary and computes present errors.
Args:
excluded_error (str): The error key to be excluded from the error dictionary.
Returns:
Tuple[List[str], Dict[str, int]]: Filtered error keys and errors with counts > 0.
"""
# Exclude a specific error and compute present errors
# Create list of error keys excluding specified error
error_keys = [k for k in self.ERRORDICT.keys() if k != excluded_error]
if not error_keys:
return error_keys, {}
# Calculate error occurrences and filter out zeros
error_counts = self.df[error_keys].sum()
errors_present = error_counts[error_counts > 0].to_dict()
return error_keys, errors_present
def _compute_unique_children(self) -> int:
"""
Computes the number of unique children based on 'Codice fiscale'.
Args:
None
Returns:
int: The number of unique children.
"""
# Compute the number of unique children
return self.df["Codice fiscale"].nunique()
def _compute_unique_gemeinden(self) -> int:
"""
Computes the number of unique municipalities based on 'Com_code'.
Args:
None
Returns:
int: The number of unique municipalities.
"""
# Compute the number of unique municipalities
return self.df["Com_code"].nunique()
def _generate_excel_report(self, nr_bambini: int, nr_comuni: int, errors_present: dict) -> bytes:
"""
Generates the Excel report and returns the binary data.
Args:
nr_bambini (int): Number of unique children.
nr_comuni (int): Number of unique municipalities.
errors_present (dict): Dictionary of errors with their counts.
Returns:
bytes: The binary data of the generated Excel report.
"""
# Generate the Excel report and return the binary data
excel_buffer = BytesIO()
with xlsxwriter.Workbook(excel_buffer) as workbook:
worksheet = workbook.add_worksheet("Riassunto")
try:
self._write_excel_summary(worksheet, nr_bambini, nr_comuni, errors_present, workbook)
except:
st.error("Error during _write_excel_summary")
try:
self._write_excel_detailed_sections(worksheet, errors_present, workbook)
except:
st.error("Error during _write_excel_detailed")
return excel_buffer.getvalue()
def _write_excel_summary(self, worksheet, nr_bambini: int, nr_comuni: int, errors_present: dict, workbook):
"""
Writes the summary section of the Excel report.
Args:
worksheet: The worksheet object to write the summary to.
nr_bambini (int): Number of unique children.
nr_comuni (int): Number of unique municipalities.
errors_present (dict): Dictionary of errors with their counts.
workbook: The workbook object to format the cells.
Returns:
None
"""
# Write the summary section of the Excel report
# Define cell formats for different sections
bold_format = workbook.add_format({'bold': True, 'font_size': 12})
bold_format2 = workbook.add_format({'bold': True, 'font_size': 11})
header_format = workbook.add_format({'bold': True, 'bg_color': '#D7E4BC'})
title_format = workbook.add_format({'bold': True, 'font_size': 18})
row, col = 3, 0 # Starting from row 1 as row 0 is typically reserved
# Write report title and basic statistics
try:
worksheet.write(row, col, "Riassunto analisi resoconti Tagesmuetter", title_format)
worksheet.set_row(row, 30) # Set row height to 30
row += 4
except xlsxwriter.exceptions.XlsxWriterException as xlse:
st.error(f"Worksheet write error: {str(xlse)}")
raise # Re-raise to abort the process
try:
# Number of children
worksheet.write(row, col, f"Sono stati elaborati i dati di {nr_bambini} bambini in {nr_comuni} comuni")
row += 2
except:
st.error("Error during Sono stati elaborati i dati di {nr_bambini} bambini in {nr_comuni} comuni")
# Document all errors found in the analysis
try:
for error_key, count in errors_present.items():
if error_key not in self.df.columns:
raise KeyError(f"Error column {error_key} not found in dataframe")
if error_key not in self.ERRORDICT:
raise ValueError(f"Undocumented error code: {error_key}")
error_description = self.ERRORDICT[error_key]
worksheet.write(
row, col,
f"Trovate {int(count)} occorrenze per il controllo {error_description}")
row += 1
row += 2 # Blank row
except:
st.error("Error during Document all errors found in the analysis")
# Document processing errors from file handling
try:
for key in self.processor.errori.keys():
#st.write(key)
worksheet.write(
row, col,
f"Errori {key}:" + (" (file non usati)" if key == "critici" else ""),
bold_format
)
row += 1
for filename in self.processor.errori[key]:
#st.write(filename)
worksheet.write(
row, col+1,
filename.split('/', 1)[1] if '/' in filename else filename,
#filename, #.split('/', 1)[1],
)
for error in self.processor.errori[key][filename]:
#st.write("entered error")
#st.write(error)
worksheet.write(
row, col+2,
error,
)
row += 1
row += 1
row += 3
except Exception as e:
st.error("Error during Document processing errors from file handling: " + str(e))
# Handle special cases: children without hour control
try:
if not self.df_bambini.empty:
worksheet.write(
row, col,
"Bambini per i quali non e' possibile eseguire il controllo superamento ore nel periodo contrattuale",
bold_format
)
row += 1
# Write column headers
for header_col, header in enumerate(self.df_bambini.columns):
worksheet.write(row, header_col, header, header_format)
row += 1
# Set fixed column widths
worksheet.set_column(0, len(self.df_bambini.columns)-1, 25)
# Write data rows
for _, r in self.df_bambini.iterrows():
for header_col, header in enumerate(self.df_bambini.columns):
value = r.get(header, "")
if isinstance(value, pd.Timestamp):
value = value.strftime("%d/%m/%Y")
worksheet.write(row, header_col, value)
row += 1
row += 3 # Blank rows
except:
st.error("Error during Handle special cases: children without hour control")
# Document discrepancies in attachment 666
if not self.df_calcolo_666.empty:
worksheet.set_column(0, len(self.df_calcolo_666.columns)-1, 25)
filtered_df_ore = self.df_calcolo_666[self.df_calcolo_666['Confronto ore'] == False]
filtered_df_bambini = self.df_calcolo_666[self.df_calcolo_666['Confronto bambini'] == False]
if not filtered_df_ore.empty:
worksheet.write(
row, col,
"Riscontrate incongruenze nel numero delle ore per l'allegato 666", bold_format)
row += 2
# Define headers
headers = ["Comune provenienza bambino", "Ore di servizio allegato 666", "Ore di assistenza report Comuni"]
for i, header in enumerate(headers):
worksheet.write(row, col + i + 1, header, header_format)
row += 1
for _, r in filtered_df_ore.iterrows():
worksheet.write(row, col + 1, r.get("Comune provenienza bambino", ""))
worksheet.write(row, col + 2, r.get("Ore di servizio allegato 666", ""))
worksheet.write(row, col + 3, r.get("Ore di assistenza report Comuni", ""))
row += 1
row +=1 # Blank rows
if not filtered_df_bambini.empty:
worksheet.write(
row, col,
"Riscontrate incongruenze nel numero bambini per l'allegato 666", bold_format)
row += 2
# Define headers
headers = ["Comune provenienza bambino", "Numero utenti allegato 666", "Numero bambini report Comuni"]
for i, header in enumerate(headers):
worksheet.write(row, col + i + 1, header, header_format)
row += 1
for _, r in filtered_df_bambini.iterrows():
worksheet.write(row, col + 1, r.get("Comune provenienza bambino", ""))
worksheet.write(row, col + 2, r.get("Numero utenti allegato 666", ""))
worksheet.write(row, col + 3, r.get("Numero bambini report Comuni", ""))
row += 1
def _write_excel_detailed_sections(self, worksheet, errors_present: dict, workbook):
"""
Writes detailed sections per error in the Excel report.
Args:
worksheet: The worksheet object to write the detailed sections to.
errors_present (dict): Dictionary of errors with their counts.
workbook: The workbook object to format the cells.
Returns:
None
"""
# Write detailed sections per error in the Excel report
title_format2 = workbook.add_format({'bold': True, 'font_size': 12})
header_format = workbook.add_format({'bold': True, 'bg_color': '#D7E4BC'})
date_format = workbook.add_format({'num_format': 'dd/mm/yyyy'})
summary_format = workbook.add_format({'italic': False, 'font_size': 10})
row, col = worksheet.dim_rowmax, 0
row += 4
for error_key, count in errors_present.items():
error_description = self.ERRORDICT.get(error_key, error_key)
worksheet.write(row, col, error_description, title_format2)
row += 2
# Headers for detailed sections
headers = [
"Cognome",
"Nome",
"Codice fiscale",
"Data di nascita",
"Comune",
"Data inizio contratto",
"Data fine contratto",
]
for i, header in enumerate(headers):
worksheet.write(row, col + i + 1, header, header_format)
row += 1
for header_col, header in enumerate(headers, start=1):
max_length = max(
self.df[header].astype(str).map(len).max()
if header in self.df.columns else 0,
len(header)
) + 2
worksheet.set_column(col + header_col, col + header_col, max_length + 5)
df_bambini = self.df[self.df[error_key] == True].sort_values(by=["Cognome", "Nome"])
for _, r in df_bambini.iterrows():
worksheet.write(row, col + 1, r.get("Cognome", ""))
worksheet.write(row, col + 2, r.get("Nome", ""))
worksheet.write(row, col + 3, r.get("Codice fiscale", ""))
data_nascita = r.get("Data di nascita", "")
if pd.notnull(data_nascita) and isinstance(data_nascita, pd.Timestamp):
worksheet.write_datetime(row, col + 4, data_nascita, date_format)
else:
worksheet.write(row, col + 4, str(data_nascita))
worksheet.write(row, col + 5, r.get("Comune", ""))
data_inizio = r.get("Data inizio contratto (o data inizio assistenza se diversa)", "")
if pd.notnull(data_inizio) and isinstance(data_inizio, pd.Timestamp):
worksheet.write_datetime(row, col + 6, data_inizio, date_format)
else:
worksheet.write(row, col + 6, str(data_inizio))
data_fine = r.get("Data fine contratto (o data fine assistenza se diversa)", "")
if pd.notnull(data_fine) and isinstance(data_fine, pd.Timestamp):
worksheet.write_datetime(row, col + 7, data_fine, date_format)
else:
worksheet.write(row, col + 7, str(data_fine))
row += 1
# Add summary line below the table
row += 1
worksheet.write(row, col, "Riepilogo:", summary_format)
#row += 1
# Collect all names and communes into a single summary line
summary_entries = []
for _, r in df_bambini.iterrows():
# Capitalize each part of potentially compound names
cognome_parts = [part.capitalize() for part in r.get("Cognome", "").split()]
nome_parts = [part.capitalize() for part in r.get("Nome", "").split()]
cognome = " ".join(cognome_parts).rstrip()
nome = " ".join(nome_parts).rstrip()
comune = r.get("Comune", "").rstrip()
summary_entries.append(f"{cognome} {nome} ({comune})")
# Join all entries with comma separators and write as a single line
summary_line = ", ".join(summary_entries)
worksheet.write(row, col + 1, summary_line, summary_format)
row += 1
row += 3
def _prepare_additional_df(self) -> pd.DataFrame:
"""
Prepares the additional DataFrame for the PDF report by renaming and dropping columns.
Args:
None
Returns:
pd.DataFrame: The prepared additional DataFrame.
"""
# Prepare the additional DataFrame for the PDF report
additional_df = self.df_bambini.copy()
column_list = list(additional_df.columns)
column_list[3] = 'Data Inizio Contratto'
column_list[4] = 'Data Fine Contratto'
additional_df.columns = column_list
position_to_delete = 5 # Adjust based on actual column index
if position_to_delete < len(additional_df.columns):
column_name = additional_df.columns[position_to_delete]
additional_df.drop(columns=column_name, inplace=True)
return additional_df
@st.fragment()
def _provide_downloads(self, excel_data: bytes): #, pdf_data: bytes):
"""
Provides download buttons for the generated Excel and PDF reports.
Args:
excel_data (bytes): The binary data of the generated Excel report.
# pdf_data (bytes): The binary data of the generated PDF report.
Returns:
None
"""
# Provide download buttons for the generated Excel and PDF reports
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
excel_file_name = f"riassuntoAutomatico_{timestamp}.xlsx"
col1, col2 = st.columns(2)
with col1:
st.download_button(
label="📥 Scarica report riassuntivo dell'analisi in formato Excel",
data=excel_data,
file_name=excel_file_name,
mime="application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
key="fileRiassuntoExcel",
)
|