Add loss functions and model architecture for time-to-event prediction

- Implemented ExponentialNLLLoss and WeibullNLLLoss in losses.py for negative log-likelihood calculations.
- Developed TabularEncoder class in model.py for encoding tabular features.
- Created DelphiFork and SapDelphi classes in model.py for time-to-event prediction using transformer architecture.
- Added data preparation scripts in prepare_data.R and prepare_data.py for processing UK Biobank data, including handling field mappings and event data extraction.

prepare_data.R

@@ -0,0 +1,26 @@
+library(data.table)
+setDTthreads(40)
+library(readr)
+field_id <- read.csv("field_id.txt", header = FALSE)
+uid <- field_id$V1
+big_path   <- "/mnt/storage/shared_data/UKBB/20230518-from-zhourong/HHdata_221103_0512.csv"
+header_dt  <- fread(big_path, nrows = 0)     # Read 0 rows => only column names
+all_names  <- names(header_dt)
+keep_names <- intersect(all_names,uid)
+ukb_disease <- fread(big_path,
+                 select     = keep_names,
+                 showProgress = TRUE)
+
+field_id <- read.csv("field_id.txt", header = FALSE)
+uid <- field_id$V1
+big_path <- "/mnt/storage/shared_data/UKBB/20230518-from-zhourong/HH_data_220812_0512.csv"
+header_dt  <- fread(big_path, nrows = 0)     # Read 0 rows => only column names
+all_names  <- names(header_dt)
+keep_names <- intersect(all_names,uid)
+ukb_others <- fread(big_path,
+                 select     = keep_names,
+                 showProgress = TRUE)
+
+# merge disease and other data by "eid"
+ukb_data <- merge(ukb_disease, ukb_others, by = "eid", all = TRUE)
+fwrite(ukb_data, "ukb_data.csv")