validerror_bioseqset.cpp

来自「ncbi源码」· C++ 代码 · 共 493 行 · 第 1/2 页

        } // else if
    } // iterate
    
    CTypeConstIterator<CMolInfo> miit(ConstBegin(seqset));
    const CMolInfo* mol_info = 0;
    
    for (; miit; ++miit) {
        if (mol_info == 0) {
            mol_info = &(*miit);
        } else if (mol_info->GetBiomol() != miit->GetBiomol() ) {
            PostErr(eDiag_Error, eErr_SEQ_PKG_InconsistentMolInfoBiomols,
                "Segmented set contains inconsistent MolInfo biomols",
                seqset);
            break;
        }
    } // for
}


void CValidError_bioseqset::ValidatePartsSet(const CBioseq_set& seqset)
{
    CSeq_inst::EMol     mol = CSeq_inst::eMol_not_set;
    CSeq_inst::EMol     seq_inst_mol;

    ITERATE ( list< CRef<CSeq_entry> >, se_list_it, seqset.GetSeq_set() ) {
        if ( (**se_list_it).IsSeq() ) {
            const CSeq_inst& seq_inst = (**se_list_it).GetSeq().GetInst();

            if ( mol == CSeq_inst::eMol_not_set  ||
                 mol == CSeq_inst::eMol_other ) {
                mol = seq_inst.GetMol();
            } else  {
                seq_inst_mol = seq_inst.GetMol();
                if ( seq_inst_mol != CSeq_inst::eMol_other) {
                    if ( seq_inst.IsNa() != CSeq_inst::IsNa(mol) ) {
                        PostErr(eDiag_Critical, eErr_SEQ_PKG_PartsSetMixedBioseqs,
                                 "Parts set contains mixture of nucleotides "
                                 "and proteins", seqset);
                        break;
                    }
                }
            }
        } else if ( (**se_list_it).IsSet() ) {
            const CBioseq_set& set = (**se_list_it).GetSet();
            const CEnumeratedTypeValues* tv = 
                CBioseq_set::GetTypeInfo_enum_EClass();
            const string& set_class_str = 
                tv->FindName(set.GetClass(), true);

            PostErr(eDiag_Error, eErr_SEQ_PKG_PartsSetHasSets,
                    "Parts set contains unwanted Bioseq-set, "
                    "its class is \"" + set_class_str + "\".", set);
            break;
        } // else if
    } // for
}


void CValidError_bioseqset::ValidatePopSet(const CBioseq_set& seqset)
{
    const CBioSource*   biosrc  = 0;
    const string        *first_taxname = 0;
    static const string influenza = "Influenza virus ";
    static const string sp = " sp. ";
    
    CTypeConstIterator<CBioseq> seqit(ConstBegin(seqset));
    for (; seqit; ++seqit) {
        
        biosrc = 0;
        
        // Will get the first biosource either from the descriptor
        // or feature.
        CTypeConstIterator<CBioSource> biosrc_it(ConstBegin(*seqit));
        if (biosrc_it) {
            biosrc = &(*biosrc_it);
        } 
        
        if (biosrc == 0)
            continue;
        
        const string& taxname = biosrc->GetOrg().GetTaxname();
        if (first_taxname == 0) {
            first_taxname = &taxname;
            continue;
        }
        
        // Make sure all the taxnames in the set are the same.
        if ( NStr::CompareNocase(*first_taxname, taxname) == 0 ) {
            continue;
        }
        
        // if the names differ issue an error with the exception of Influenza
        // virus, where we allow different types of it in the set.
        if ( NStr::StartsWith(taxname, influenza, NStr::eNocase)         &&
            NStr::StartsWith(*first_taxname, influenza, NStr::eNocase)  &&
            NStr::CompareNocase(*first_taxname, 0, influenza.length() + 1, taxname) == 0 ) {
            continue;
        }
        
        // drops severity if first mismatch is same up to sp.
        EDiagSev sev = eDiag_Error;
        SIZE_TYPE pos = NStr::Find(taxname, sp);
        if ( pos != NPOS ) {
            SIZE_TYPE len = pos + sp.length();
            if ( NStr::strncasecmp(first_taxname->c_str(), 
                                   taxname.c_str(),
                                   len) == 0 ) {
                sev = eDiag_Warning;
            }
        }

        PostErr(eDiag_Error, eErr_SEQ_DESCR_InconsistentBioSources,
            "Population set contains inconsistent organisms.",
            *seqit);
        break;
    }
}


void CValidError_bioseqset::ValidateGenProdSet(const CBioseq_set& seqset)
{
    bool                id_no_good = false;
    CSeq_id::E_Choice   id_type = CSeq_id::e_not_set;
    
    list< CRef<CSeq_entry> >::const_iterator se_list_it =
        seqset.GetSeq_set().begin();
    
    if ( !(**se_list_it).IsSeq() ) {
        return;
    }
    
    const CBioseq& seq = (*se_list_it)->GetSeq();
    CBioseq_Handle bsh = m_Scope->GetBioseqHandle(seq);

    CFeat_CI fi(bsh, 0, 0, CSeqFeatData::e_Rna);
    for (; fi; ++fi) {
        if ( fi->GetData().GetSubtype() == CSeqFeatData::eSubtype_mRNA ) {
            if ( fi->IsSetProduct() ) {
                CBioseq_Handle cdna = 
                    m_Scope->GetBioseqHandle(fi->GetProduct());
                if ( !cdna ) {
                    try {
                        const CSeq_id& id = GetId(fi->GetProduct(), m_Scope);
                        id_type = id.Which();
                    }
                    catch (...) {
                        id_no_good = true;
                    }
                    
                    // okay to have far RefSeq product
                    if ( id_no_good  ||  (id_type != CSeq_id::e_Other) ) {
                        string loc_label;
                        fi->GetProduct().GetLabel(&loc_label);
                        
                        if (loc_label.empty()) {
                            loc_label = "?";
                        }
                        
                        PostErr(eDiag_Error,
                            eErr_SEQ_PKG_GenomicProductPackagingProblem,
                            "Product of mRNA feature (" + loc_label +
                            ") not packaged in genomic product set", seq);
                        
                    }
                } // if (cdna == 0)
            } else {
                PostErr(eDiag_Error,
                    eErr_SEQ_PKG_GenomicProductPackagingProblem,
                    "Product of mRNA feature (?) not packaged in"
                    "genomic product set", seq);
            }
        }
    } // for 
}


END_SCOPE(validator)
END_SCOPE(objects)
END_NCBI_SCOPE


/*
* ===========================================================================
*
* $Log: validerror_bioseqset.cpp,v $
* Revision 1000.2  2004/06/01 19:47:54  gouriano
* PRODUCTION: UPGRADED [GCC34_MSVC7] Dev-tree R1.14
*
* Revision 1.14  2004/05/21 21:42:56  gorelenk
* Added PCH ncbi_pch.hpp
*
* Revision 1.13  2004/04/05 15:56:15  grichenk
* Redesigned CAnnotTypes_CI: moved all data and data collecting
* functions to CAnnotDataCollector. CAnnotTypes_CI is no more
* inherited from SAnnotSelector.
*
* Revision 1.12  2003/06/02 16:06:43  dicuccio
* Rearranged src/objects/ subtree.  This includes the following shifts:
*     - src/objects/asn2asn --> arc/app/asn2asn
*     - src/objects/testmedline --> src/objects/ncbimime/test
*     - src/objects/objmgr --> src/objmgr
*     - src/objects/util --> src/objmgr/util
*     - src/objects/alnmgr --> src/objtools/alnmgr
*     - src/objects/flat --> src/objtools/flat
*     - src/objects/validator --> src/objtools/validator
*     - src/objects/cddalignview --> src/objtools/cddalignview
* In addition, libseq now includes six of the objects/seq... libs, and libmmdb
* replaces the three libmmdb? libs.
*
* Revision 1.11  2003/04/29 14:45:47  shomrat
* Bug fix in ValidateSegSet
*
* Revision 1.10  2003/03/31 14:40:42  shomrat
* $id: -> $id$
*
* Revision 1.9  2003/03/18 21:48:37  grichenk
* Removed obsolete class CAnnot_CI
*
* Revision 1.8  2003/03/11 16:04:09  kuznets
* iterate -> ITERATE
*
* Revision 1.7  2003/02/14 15:11:57  shomrat
* population study check for inconsistent organisms drops severity if first mismatch is same up to sp.
*
* Revision 1.6  2003/02/07 21:18:38  shomrat
* Initialize variable
*
* Revision 1.5  2003/01/30 20:27:20  shomrat
* Bug fix
*
* Revision 1.4  2003/01/06 17:37:19  shomrat
* Fixed mismatch types
*
* Revision 1.3  2003/01/02 21:54:33  shomrat
* Implemented IsMrnaProductInGPS
*
* Revision 1.2  2002/12/24 16:53:24  shomrat
* Changes to include directives
*
* Revision 1.1  2002/12/23 20:16:16  shomrat
* Initial submission after splitting former implementation
*
*
* ===========================================================================
*/