Mercurial > repos > dchristiany > data_manager_proteore

comparison data_manager/resource_building.py @ 12:60cb0a5ae661 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload commit e5e768b479ddc6b36270a1b5b0443a4c80d693bc-dirty
author dchristiany
date Tue, 23 Oct 2018 08:18:32 -0400
parents 2f153b41b6fe
children a1530507fee4
comparison
equal deleted inserted replaced
11:4bb219eee10f 12:60cb0a5ae661
50 unzip(url, path) #download and save file 50 unzip(url, path) #download and save file
51 tissue_name = tissue_name + " " + time.strftime("%d/%m/%Y") 51 tissue_name = tissue_name + " " + time.strftime("%d/%m/%Y")
52 tissue_id = tissue_name.replace(" ","_").replace("/","-") 52 tissue_id = tissue_name.replace(" ","_").replace("/","-")
53 53
54 data_table_entry = dict(id=tissue_id, name = tissue_name, value = tissue, path = path) 54 data_table_entry = dict(id=tissue_id, name = tissue_name, value = tissue, path = path)
55 _add_data_table_entry(data_manager_dict, data_table_entry, "protein_atlas") 55 _add_data_table_entry(data_manager_dict, data_table_entry, "proteore_protein_atlas")
56 56
57 57
58 ####################################################################################################### 58 #######################################################################################################
59 # 2. Peptide Atlas 59 # 2. Peptide Atlas
60 ####################################################################################################### 60 #######################################################################################################
88 w = csv.writer(out,delimiter='\t') 88 w = csv.writer(out,delimiter='\t')
89 w.writerow(["Uniprot_AC","nb_obs"]) 89 w.writerow(["Uniprot_AC","nb_obs"])
90 w.writerows(uni_dict.items()) 90 w.writerows(uni_dict.items())
91 91
92 data_table_entry = dict(id=tissue_id, name=tissue_name, value = path, tissue = tissue) 92 data_table_entry = dict(id=tissue_id, name=tissue_name, value = path, tissue = tissue)
93 _add_data_table_entry(data_manager_dict, data_table_entry, "peptide_atlas") 93 _add_data_table_entry(data_manager_dict, data_table_entry, "proteore_peptide_atlas")
94 94
95 #function to count the number of observations by uniprot id 95 #function to count the number of observations by uniprot id
96 def build_dictionary (csv) : 96 def build_dictionary (csv) :
97 uni_dict = {} 97 uni_dict = {}
98 for line in csv : 98 for line in csv :
219 name_dict={"human" : "Homo sapiens", "mouse" : "Mus musculus", "rat" : "Rattus norvegicus"} 219 name_dict={"human" : "Homo sapiens", "mouse" : "Mus musculus", "rat" : "Rattus norvegicus"}
220 name = name_dict[species]+" "+time.strftime("%d/%m/%Y") 220 name = name_dict[species]+" "+time.strftime("%d/%m/%Y")
221 id = species+"_id_mapping_"+ time.strftime("%d-%m-%Y") 221 id = species+"_id_mapping_"+ time.strftime("%d-%m-%Y")
222 222
223 data_table_entry = dict(id=id, name = name, value = species, path = path) 223 data_table_entry = dict(id=id, name = name, value = species, path = path)
224 _add_data_table_entry(data_manager_dict, data_table_entry, "id_mapping") 224 _add_data_table_entry(data_manager_dict, data_table_entry, "proteore_id_mapping")
225 225
226 def download_from_uniprot_ftp(file,target_directory) : 226 def download_from_uniprot_ftp(file,target_directory) :
227 ftp_dir = "pub/databases/uniprot/current_release/knowledgebase/idmapping/by_organism/" 227 ftp_dir = "pub/databases/uniprot/current_release/knowledgebase/idmapping/by_organism/"
228 path = os.path.join(target_directory, file) 228 path = os.path.join(target_directory, file)
229 ftp = ftplib.FTP("ftp.uniprot.org") 229 ftp = ftplib.FTP("ftp.uniprot.org")