20230531_GDAD_initiation

.vscode/launch.json

@@ -0,0 +1,19 @@
+{
+    // 使用 IntelliSense 了解相关属性。 
+    // 悬停以查看现有属性的描述。
+    // 欲了解更多信息，请访问: https://go.microsoft.com/fwlink/?linkid=830387
+    "version": "0.2.0",
+    "configurations": [
+        {
+            "name": "Python: Django",
+            "type": "python",
+            "request": "launch",
+            "program": "${workspaceFolder}\\mysite\\manage.py",
+            "args": [
+                "runserver"
+            ],
+            "django": true,
+            "justMyCode": true
+        }
+    ]
+}

mysite/gdad/admin.py

@@ -0,0 +1,10 @@
+from django.contrib import admin
+
+# Register your models here.
+
+from .models import G4, GeneSequence, Tfbs, Gda
+
+admin.site.register(G4)
+admin.site.register(GeneSequence)
+admin.site.register(Tfbs)
+admin.site.register(Gda)

mysite/gdad/apps.py

@@ -0,0 +1,6 @@
+from django.apps import AppConfig
+
+
+class GdadConfig(AppConfig):
+    default_auto_field = "django.db.models.BigAutoField"
+    name = "gdad"

mysite/gdad/migrations/0001_initial.py

@@ -0,0 +1,103 @@
+# Generated by Django 4.1.7 on 2023-03-14 11:45
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+    initial = True
+
+    dependencies = []
+
+    operations = [
+        migrations.CreateModel(
+            name="G4",
+            fields=[
+                (
+                    "id",
+                    models.BigAutoField(
+                        auto_created=True,
+                        primary_key=True,
+                        serialize=False,
+                        verbose_name="ID",
+                    ),
+                ),
+                ("chr", models.CharField(default="", max_length=31)),
+                ("start", models.PositiveIntegerField()),
+                ("end", models.PositiveIntegerField()),
+                ("strand", models.CharField(default="", max_length=1)),
+                ("cell_line", models.CharField(default="", max_length=63)),
+                ("technology", models.CharField(default="", max_length=63)),
+                ("doi", models.CharField(default="", max_length=255)),
+            ],
+        ),
+        migrations.CreateModel(
+            name="Gda",
+            fields=[
+                (
+                    "id",
+                    models.BigAutoField(
+                        auto_created=True,
+                        primary_key=True,
+                        serialize=False,
+                        verbose_name="ID",
+                    ),
+                ),
+                ("gene_id", models.CharField(default="", max_length=31)),
+                ("gene_symbol", models.CharField(default="", max_length=31)),
+                ("uniprot_id", models.CharField(default="", max_length=63)),
+                ("gene_dsi", models.FloatField()),
+                ("gene_dpi", models.FloatField()),
+                ("gene_pli", models.FloatField()),
+                ("protein_class", models.CharField(default="", max_length=31)),
+                ("protein_class_name", models.CharField(default="", max_length=63)),
+                ("disease_id", models.CharField(default="", max_length=15)),
+                ("disease_name", models.CharField(default="", max_length=255)),
+                ("disease_class", models.CharField(default="", max_length=255)),
+                ("disease_class_name", models.CharField(default="", max_length=1023)),
+                ("disease_type", models.CharField(default="", max_length=15)),
+                ("disease_semantic_type", models.CharField(default="", max_length=255)),
+                ("gda_score", models.FloatField()),
+                ("ei", models.FloatField()),
+                ("el", models.CharField(default="", max_length=31)),
+                ("year_initial", models.PositiveIntegerField()),
+                ("year_final", models.PositiveIntegerField()),
+            ],
+        ),
+        migrations.CreateModel(
+            name="GeneSequence",
+            fields=[
+                (
+                    "id",
+                    models.BigAutoField(
+                        auto_created=True,
+                        primary_key=True,
+                        serialize=False,
+                        verbose_name="ID",
+                    ),
+                ),
+                ("chr", models.CharField(default="", max_length=31)),
+                ("start", models.PositiveIntegerField()),
+                ("end", models.PositiveIntegerField()),
+                ("gene_symbol", models.CharField(default="", max_length=31)),
+            ],
+        ),
+        migrations.CreateModel(
+            name="Tfbs",
+            fields=[
+                (
+                    "id",
+                    models.BigAutoField(
+                        auto_created=True,
+                        primary_key=True,
+                        serialize=False,
+                        verbose_name="ID",
+                    ),
+                ),
+                ("chr", models.CharField(default="", max_length=31)),
+                ("start", models.PositiveIntegerField()),
+                ("end", models.PositiveIntegerField()),
+                ("tf", models.CharField(default="", max_length=31)),
+                ("cell_line", models.CharField(default="", max_length=1023)),
+            ],
+        ),
+    ]

mysite/gdad/migrations/0002_alter_gda_gene_pli.py

@@ -0,0 +1,17 @@
+# Generated by Django 4.1.7 on 2023-03-22 02:24
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("gdad", "0001_initial"),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name="gda",
+            name="gene_pli",
+            field=models.CharField(default="", max_length=15),
+        ),
+    ]

mysite/gdad/models.py

@@ -0,0 +1,164 @@
+from django.db import models
+
+# Create your models here.
+
+
+class G4(models.Model):
+    # g4
+    chr = models.CharField(max_length=31, default="")
+    start = models.PositiveIntegerField()
+    end = models.PositiveIntegerField()
+    strand = models.CharField(max_length=1, default="")
+    cell_line = models.CharField(max_length=63, default="")
+    technology = models.CharField(max_length=63, default="")
+    doi = models.CharField(max_length=255, default="")
+
+    def __str__(self):
+        out_list = [
+            self.chr,
+            self.start,
+            self.end,
+            self.strand,
+            self.cell_line,
+            self.technology,
+            self.doi,
+        ]
+        out_list = list(map(lambda x: str(x), out_list))
+        out_str = "\t".join(out_list)
+        return out_str
+
+    def get_field_list(self) -> list[str]:
+        """打印字段名列表。
+
+        Returns:
+            field_list (list[str]): 字段名列表。
+        """
+        field_list = [
+            "chr",
+            "start",
+            "end",
+            "strand",
+            "cell_line",
+            "technology",
+            "doi",
+        ]
+        return field_list
+
+
+class GeneSequence(models.Model):
+    # gene_sequence
+    chr = models.CharField(max_length=31, default="")
+    start = models.PositiveIntegerField()
+    end = models.PositiveIntegerField()
+    gene_symbol = models.CharField(max_length=31, default="")
+
+    def __str__(self):
+        out_list = [self.chr, self.start, self.end, self.gene_symbol]
+        out_list = list(map(lambda x: str(x), out_list))
+        out_str = "\t".join(out_list)
+        return out_str
+
+    def get_field_list(self) -> list[str]:
+        """打印字段名列表。
+
+        Returns:
+            field_list (list[str]): 字段名列表。
+        """
+        field_list = [
+            "chr",
+            "start",
+            "end",
+            "gene_symbol",
+        ]
+        return field_list
+
+
+class Tfbs(models.Model):
+    # tfbs，没啥用
+    chr = models.CharField(max_length=31, default="")
+    start = models.PositiveIntegerField()
+    end = models.PositiveIntegerField()
+    tf = models.CharField(max_length=31, default="")
+    cell_line = models.CharField(max_length=1023, default="")
+
+    def __str__(self):
+        out_list= [self.chr, self.start, self.end, self.tf, self.cell_line]
+        out_list = list(map(lambda x: str(x), out_list))
+        out_str="\t".join(out_list)
+        return out_str
+
+    def get_field_list(self) -> list[str]:
+        """打印字段名列表。
+
+        Returns:
+            field_list (list[str]): 字段名列表。
+        """
+        field_list = [
+            "chr",
+            "start",
+            "end",
+            "tf",
+            "cell_line",
+        ]
+        return field_list
+
+
+class Gda(models.Model):
+    # gda
+    gene_id = models.CharField(max_length=31, default="")
+    gene_symbol = models.CharField(max_length=31, default="")
+    uniprot_id = models.CharField(max_length=63, default="")
+    gene_dsi = models.FloatField()
+    gene_dpi = models.FloatField()
+    # gene_pli字段有很多科学计数法，差距过大，数值型字段无法支持
+    gene_pli = models.CharField(max_length=15, default="")
+    protein_class = models.CharField(max_length=31, default="")
+    protein_class_name = models.CharField(max_length=63, default="")
+    disease_id = models.CharField(max_length=15, default="")
+    disease_name = models.CharField(max_length=255, default="")
+    disease_class = models.CharField(max_length=255, default="")
+    disease_class_name = models.CharField(max_length=1023, default="")
+    disease_type = models.CharField(max_length=15, default="")
+    disease_semantic_type = models.CharField(max_length=255, default="")
+    gda_score = models.FloatField()
+    ei = models.FloatField()
+    el = models.CharField(max_length=31, default="")
+    year_initial = models.PositiveIntegerField()
+    year_final = models.PositiveIntegerField()
+
+    def __str__(self):
+        out_list=[self.gene_id,
+                  self.gene_symbol,self.disease_id,self.disease_name,
+                  ]
+        out_list=list(map(lambda x:str(x),out_list))
+        out_str="\t".join(out_list)
+        return out_str
+
+    def get_field_list(self) -> list[str]:
+        """打印字段名列表。
+
+        Returns:
+            field_list (list[str]): 字段名列表。
+        """
+        field_list = [
+            "gene_id",
+            "gene_symbol",
+            "uniprot_id",
+            "gene_dsi",
+            "gene_dpi",
+            "gene_pli",
+            "protein_class",
+            "protein_class_name",
+            "disease_id",
+            "disease_name",
+            "disease_class",
+            "disease_class_name",
+            "disease_type",
+            "disease_semantic_type",
+            "gda_score",
+            "ei",
+            "el",
+            "year_initial",
+            "year_final",
+        ]
+        return field_list