|
1 | | -import gzip |
2 | | -import json |
3 | | -import os |
4 | | -from collections import defaultdict |
5 | | -from typing import Optional |
6 | | - |
7 | 1 | from flask import jsonify, make_response |
8 | 2 |
|
9 | | -from backend.cellguide.common.config import CellGuideConfig |
10 | | -from backend.cellguide.common.constants import COMPUTATIONAL_MARKER_GENES_FOLDERNAME, MARKER_GENE_PRESENCE_FILENAME |
11 | | -from backend.cellguide.common.providers.s3_provider import S3Provider |
12 | | -from backend.cellguide.common.utils import get_object_key |
13 | | -from backend.common.utils.ontology_parser import ontology_parser |
14 | | - |
15 | | - |
16 | | -def _defaultdict_to_dict(d): |
17 | | - if isinstance(d, defaultdict): |
18 | | - # Convert the defaultdict to a dict and recursively apply this function |
19 | | - return {key: _defaultdict_to_dict(value) for key, value in d.items()} |
20 | | - else: |
21 | | - return d |
22 | | - |
23 | | - |
24 | | -def _initialize_cellguide_marker_gene_dict(): |
25 | | - bucket = CellGuideConfig().bucket |
26 | | - s3_provider = S3Provider() |
27 | | - |
28 | | - latest_snapshot_identifier = ( |
29 | | - s3_provider.download_file(bucket_name=bucket, object_key=get_object_key(object="latest_snapshot_identifier")) |
30 | | - .decode("utf-8") |
31 | | - .strip() |
32 | | - ) |
33 | | - compressed_data = s3_provider.download_file( |
34 | | - bucket_name=bucket, |
35 | | - object_key=get_object_key( |
36 | | - object=f"{latest_snapshot_identifier}/{COMPUTATIONAL_MARKER_GENES_FOLDERNAME}/{MARKER_GENE_PRESENCE_FILENAME}" |
37 | | - ), |
38 | | - ) |
39 | | - marker_gene_data = json.loads(gzip.decompress(compressed_data).decode("utf-8")) |
40 | | - data = defaultdict(lambda: defaultdict(lambda: defaultdict(list))) |
41 | | - |
42 | | - for gene in marker_gene_data: |
43 | | - for organism in marker_gene_data[gene]: |
44 | | - for tissue in marker_gene_data[gene][organism]: |
45 | | - for marker in marker_gene_data[gene][organism][tissue]: |
46 | | - data[organism][tissue][marker["cell_type_id"]].append( |
47 | | - {"marker_score": marker["marker_score"], "me": marker["me"], "pc": marker["pc"], "gene": gene} |
48 | | - ) |
49 | | - |
50 | | - data = _defaultdict_to_dict(data) |
51 | | - |
52 | | - for organism in data: |
53 | | - for tissue in data[organism]: |
54 | | - for cell_type in data[organism][tissue]: |
55 | | - data[organism][tissue][cell_type].sort(key=lambda x: -x["marker_score"]) |
56 | | - |
57 | | - return data |
58 | | - |
59 | | - |
60 | | -def is_id(term: str) -> bool: |
61 | | - return term.startswith("CL:") or term.startswith("UBERON:") or term.startswith("NCBITaxon:") |
| 3 | +from backend.cellguide.api.common.data import get_marker_gene_data |
62 | 4 |
|
63 | 5 |
|
64 | | -def get(organism: str, tissue: Optional[str] = None, cell_type: Optional[str] = None): |
| 6 | +def get(): |
65 | 7 | """ |
66 | | - Retrieve marker gene data for a specified organism, and optionally for a specific tissue and/or cell type. |
| 8 | + Retrieve all CellGuide marker gene data. |
67 | 9 |
|
68 | | - This function handles the retrieval of marker gene data based on the provided organism, tissue, and cell type. |
69 | | - It uses ontology IDs to resolve the actual names and checks if the provided IDs exist in the marker gene data. |
70 | | - If the IDs are valid, it returns the corresponding data; otherwise, it returns a 404 response. |
71 | | -
|
72 | | - Parameters: |
73 | | - organism (str): The ontology ID or name of the organism. |
74 | | - tissue (Optional[str]): The ontology ID or name of the tissue. Default is None. |
75 | | - cell_type (Optional[str]): The ontology ID or name of the cell type. Default is None. |
| 10 | + This function handles the retrieval of all CellGuide marker gene data. |
76 | 11 |
|
77 | 12 | Returns: |
78 | | - Flask Response: JSON data of the marker genes for the requested parameters or a 404 error if not found. |
| 13 | + Flask Response: JSON data of the marker genes. |
| 14 | + The response structure is a nested dictionary with the following structure: |
| 15 | + - organisms --> tissues --> cell types --> list(marker genes). |
| 16 | +
|
| 17 | + Organisms and tissues are labels, cell types are IDs, and marker genes are dictionaries with the following keys: |
| 18 | + - `marker_score`: The score indicating the strength of the marker gene for the cell type. |
| 19 | + - `me`: Mean expression of the gene across the cells of the specified type. |
| 20 | + - `pc`: Percentage of cells within the specified type that express the gene. |
| 21 | + - `gene`: The gene symbol associated with the marker gene data. |
79 | 22 | """ |
80 | | - marker_gene_data = get_marker_gene_data() |
81 | | - |
82 | | - if is_id(organism): |
83 | | - organism = ontology_parser.get_term_label(organism) |
84 | | - if tissue and is_id(tissue): |
85 | | - tissue = ontology_parser.get_term_label(tissue) |
86 | | - if tissue and cell_type and not is_id(cell_type): |
87 | | - raise ValueError("cell_type must be an ID") |
88 | | - |
89 | | - if not tissue and cell_type or tissue and tissue.lower() == "all tissues": |
90 | | - tissue = "All Tissues" |
91 | | - |
92 | | - if organism not in marker_gene_data: |
93 | | - return make_response(jsonify({}), 404) |
94 | | - |
95 | | - if tissue and tissue not in marker_gene_data[organism]: |
96 | | - return make_response(jsonify({}), 404) |
97 | | - |
98 | | - if tissue and cell_type and cell_type not in marker_gene_data[organism][tissue]: |
99 | | - return make_response(jsonify({}), 404) |
100 | | - |
101 | | - if tissue and cell_type: |
102 | | - return make_response(jsonify(marker_gene_data[organism][tissue][cell_type]), 200) |
103 | | - elif tissue: |
104 | | - return make_response(jsonify(marker_gene_data[organism][tissue]), 200) |
105 | | - else: |
106 | | - return make_response(jsonify(marker_gene_data[organism]), 200) |
107 | | - |
108 | | - |
109 | | -_marker_gene_data_cache = None |
110 | | - |
111 | | - |
112 | | -def get_marker_gene_data(): |
113 | | - global _marker_gene_data_cache |
114 | | - if _marker_gene_data_cache is None: |
115 | | - if os.getenv("DEPLOYMENT_STAGE") != "test": |
116 | | - # Initialize the marker gene data from the latest snapshot only if not in test mode |
117 | | - _marker_gene_data_cache = _initialize_cellguide_marker_gene_dict() |
118 | | - else: |
119 | | - # Initialize an empty structure if in test mode |
120 | | - _marker_gene_data_cache = {} |
121 | | - return _marker_gene_data_cache |
| 23 | + return make_response(jsonify(get_marker_gene_data()), 200) |
0 commit comments