Mercurial > repos > chrisd > testshed

annotate gene_fraction/src/FastaRecord.h @ 0:0fd352f62446 draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
author chrisd
date Sun, 21 Feb 2016 06:05:24 -0500
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
1 #ifndef FASTA_RECORD_H
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
2 #define FASTA_RECORD_H
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
3
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
4 #include <string>
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
5 #include <vector>
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
6
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
7 /**
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
8 * Class for dealing with fasta records
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
9 */
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
10 class FastaRecord {
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
11 public:
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
12 /**
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
13 * Ctor that initializes gene id and gene
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
14 */
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
15 FastaRecord(std::string gene_id, std::string gene);
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
16
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
17 /**
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
18 * Returns a string gene id
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
19 */
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
20 std::string gene_id() const;
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
21
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
22 /**
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
23 * Returns the gene associated with gene id
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
24 */
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
25 std::string gene() const;
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
26
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
27 /**
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
28 * Returns the total base hits for a gene
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
29 */
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
30 int get_base_hits() const;
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
31
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
32 /**
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
33 * Returns the amount of genes that were hit
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
34 * during the gene fraction calculation
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
35 */
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
36 int gene_hits() const;
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
37
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
38 /**
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
39 *
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
40 */
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
41 void update_base_hits(const int &index);
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
42
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
43 /**
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
44 *
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
45 */
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
46 void update_gene_hits();
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
47
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
48 /**
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
49 * Searches for a fasta record corresponding
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
50 * to gene id
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
51 */
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
52 static int find_gene(const std::vector<FastaRecord> &records,
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
53 const std::string &gene_id,
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
54 std::string seq = "");
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
55
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
56 /**
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
57 * Sorts fasta records by gene id
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
58 */
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
59 static void sort_by_gene_id(std::vector<FastaRecord> &records);
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
60
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
61 /**
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
62 * Resets base hits vector to 0's.
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
63 * This occurs after each sample is processed
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
64 */
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
65 static void reset_base_hits(std::vector<FastaRecord> &records);
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
66
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
67 /**
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
68 * Resets gene hits primitive to 0.
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
69 * This happens after each sample is processed
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
70 */
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
71 static void reset_gene_hits(std::vector<FastaRecord> &records);
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
72
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
73 std::vector<int> &base_hits();
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
74
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
75 std::string _gene_id;
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
76 std::string _gene;
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
77 std::vector<int> _base_hits;
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
78
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
79 private:
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
80 int _gene_hits;
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
81 };
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
82
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
83
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
84
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
85
0fd352f62446 planemo upload for repository https://github.com/ChrisD11/Duplicon commit 3ee0594c692faac542ffa58f4339d79b9b8aefbd-dirty
chrisd
parents:
diff changeset
86 #endif /* FASTA_RECORD_H */