o
Rŀgj @ sp d Z ddlZddlmZ ddlmZ ddlmZ ddlmZ ddlm Z G dd d eZ
G d
d d e
Z G d d
d
e Z G dd de
Z
G dd de
ZG dd de
ZG dd de
ZG dd de
ZG dd de
ZG dd de
ZG d d d e
Zi d e
d ed e
d!ed"ed#ed$ed%ed&ed'ed(ed)e
d*e
d+e d,e d-ed.ee
ed/ZdS )0aO Dictionary like indexing of sequence files (PRIVATE).
You are not expected to access this module, or any of its code, directly. This
is all handled internally by the Bio.SeqIO.index(...) and index_db(...)
functions which are the public interface for this functionality.
The basic idea is that we scan over a sequence file, looking for new record
markers. We then try to extract the string that Bio.SeqIO.parse/read would
use as the record id, ideally without actually parsing the full record. We
then use a subclassed Python dictionary to record the file offset for the
record start against the record id.
Note that this means full parsing is on demand, so any invalid or problem
record may not trigger an exception until it is accessed. This is by design.
This means our dictionary like objects have in memory ALL the keys (all the
record identifiers), which shouldn't be a problem even with second generation
sequencing. If memory is an issue, the index_db(...) interface stores the
keys and offsets in an SQLite database - which can be re-used to avoid
re-indexing the file for use another time.
N)BytesIO)StringIO)SeqIO)_IndexedSeqFileProxy)_open_for_random_accessc @ e Zd ZdZdd Zdd ZdS )SeqFileRandomAccessz8Base class for defining random access to sequence files.c C s t || _|| _tj| | _dS Initialize the class.N)r _handle_formatr _FormatToIterator _iteratorselffilenameformat r D/var/www/html/myenv/lib/python3.10/site-packages/Bio/SeqIO/_index.py__init__) s
z SeqFileRandomAccess.__init__c C s t | t| | S )zReturn SeqRecord.)nextr r get_rawdecode)r offsetr r r get1 s zSeqFileRandomAccess.getN)__name__
__module__ __qualname____doc__r r r r r r r & s r c @ s0 e Zd ZdZdd Zdd Zdd Zdd Zd
S ) SffRandomAccessz7Random access to a Standard Flowgram Format (SFF) file.c C s4 t | || tj| j\}}}}| _| _| _dS r ) r r r SffIO_sff_file_headerr _flows_per_read _flow_chars
_key_sequence)r r r
header_length index_offset index_lengthnumber_of_readsr r r r A s zSffRandomAccess.__init__c
c s | j }|d tj|\}}}}| _| _| _|r|rd}d}z(tj|D ]\}} t || }|| dfV |d7 }q(||krHt
d||f W n2 t
y{ }
z&ddl } ddl m
} | d|
| |dkslJ d|d W Y d}
~
n"d}
~
ww || |kr|| tj|| j tj||| dS d}tj|D ]\}} || dfV |d7 }q||krt
d||f tj||| dS )zELoad any index block in the file, or build it the slow way (PRIVATE).r z Indexed %i records, expected %iN)BiopythonParserWarningz Could not parse the SFF index: zPartially populated index)r seekr r r! r" r# r$ _sff_read_roche_indexmax
ValueErrorwarningsBior* warn_sff_read_raw_record
_check_eof_sff_do_slow_index)
r handler% r& r' r( count
max_offsetnamer errr/ r* r r r __iter__N sb
zSffRandomAccess.__iter__c C s( | j }|| tj|| j| j| jS )2Return the SeqRecord starting at the given offset.r r+ r r _sff_read_seq_recordr" r# r$ r r r5 r r r r s
zSffRandomAccess.getc C s | j }|| tj|| jS 6Return the raw record from the file as a bytes string.)r r+ r r r2 r" r> r r r r s
zSffRandomAccess.get_rawN)r r r r r r: r r r r r r r > s
> r c @ e Zd ZdZdd ZdS )SffTrimedRandomAccesszLRandom access to an SFF file with defined trimming applied to each sequence.c C s, | j }|| tjj|| j| j| jddS )r; T)trimr< r> r r r r s
zSffTrimedRandomAccess.getN)r r r r r r r r r rB rB c @ ( e Zd ZdZdd Zdd Zdd ZdS ) SequentialSeqFileRandomAccessz3Random access to a simple sequential sequence file.c C sH t | || dddddddddddd | }|| _td | | _d
S ) r
s CO s ID >s LOCUS s BEGIN_SEQUENCEs >..;s s TN line in bytes %i to %i) r r+ rW rZ r[ r\ rY findr^ r. r r_ ) r r5 ra start_acc_markerend_acc_markerend_entry_markerrb rc re rw rf r r r r: sZ
"
"z UniprotRandomAccess.__iter__c C s | j }| j}d}|| | g} | }||}|dkr,||d|d n||s3|s7td|| qd|S )r@ r Tr Nr r rg ) r rW r+ r[ r rh r\ r. ri )r r r5 ra r datarc ir r r r s
zUniprotRandomAccess.get_rawc C s& d| | d }ttjt|S )r; s
s
)r r r UniprotIOUniprotIteratorr )r r r r r r r s
zUniprotRandomAccess.getN)r r r r r: r r r r r r r s
3 r c @ rE ) IntelliGeneticsRandomAccessz(Random access to a IntelliGenetics file.c C s t | || td| _dS )r
s ^;N)r r rU rV rW r r r r r s z$IntelliGeneticsRandomAccess.__init__c c s | j }|d d}d} |t|7 }| }|sn|ds!nq|rd}|t| | ks2J |ds>td||drR|t|7 }| }|dsC| }|rn|dsn|t|7 }| }|rn|dr]| ||fV ||7 }|t| | ksJ |s$dS dS )rn r Ts ;;r} z'Records should start with ';' and not:
N) r r+ rY r[ rq rZ r. rr r )r r5 r rc re rw r r r r: s>
z$IntelliGeneticsRandomAccess.__iter__c C s | j }|| | j}g }| }|dr$|| | }|ds|r;|ds;|| | }|r;|dr+d|S )r@ r} rg )r r+ rW r[ rq rh ri rj r r r r s
z#IntelliGeneticsRandomAccess.get_rawNrl r r r r r s
r c @ r )TabRandomAccessz&Random access to a simple tabbed file.c c sv | j }|d d} | }| }|sdS z ||d }W n ty/ | s.Y q w | |t|fV q )rn r TN) r r+ rZ r[ r^ r. r] r rY )r r5 tab_charrb rc rw r r r r: ( s$
zTabRandomAccess.__iter__c C s | j }|| | S r? )r r+ r[ r> r r r r = s
zTabRandomAccess.get_rawNr r r r r: r r r r r r % s r c @ r )FastqRandomAccesszRandom access to a FASTQ file (any supported variant).
With FASTQ the records all start with a "@" line, but so can quality lines.
Note this will cope with line-wrapped FASTQ files.
c c s | j }|d d}| }| }|sdS |dd dkr&td||r|dd ddd }d}t|}|rX| }|t|7 }|drNn
|t| 7 }|s>|s^tdd}|r||kr|dkr| }| rytd||t|7 }| }| }|r|dd dkrtd |n| }|t| 7 }|t|7 }|sb||krtd
|
||fV |}|s(dS dS ) rn r Nr) @Problem with FASTQ @ line:
+$Premature end of file in seq section!Expected blank quality line, not Problem with line Problem with quality section) r r+ rZ r[ r. rr r^ rY rq r] r ) r r5 rd rb rc seq_lenre qual_lenrf r r r r: P s^
zFastqRandomAccess.__iter__c C s4 | j }|| | }|}|dd dkr td|d}|r9| }||7 }|dr/n
|t| 7 }|s!|s?td|dd dksIJ d}|r||kr~|dkrh| }| rdtd|||7 }| }|r}|dd dkr}td|n| }||7 }|t| 7 }|sM||krtd |S )
r@ r r) r r r r r r r )r r+ r[ r. rq rY r] )r r r5 rc r r r r r r r sH
zFastqRandomAccess.get_rawNr r r r r r I s 7r rH rI rJ fastqz fastq-sangerz fastq-solexazfastq-illuminarK rL igrM rN rO sffzsff-trimrQ tab)rP rR )r rU ior r r0 r Bio.Filer r r r rB rF rm rz r r r r r _FormatToRandomAccessr r r r sp ]I>:"[7$l