def apropos(keyword="", notes=False):
"""
Search in all command description files those related to a user-defined keyword.
"""
out_list = set()
for i in CmdManager.cmd_obj_list:
if keyword in CmdManager.cmd_obj_list[i].desc:
out_list.add(i)
if notes:
try:
if keyword in CmdManager.cmd_obj_list[i].notes:
out_list.add(i)
except:
pass
if out_list:
message(">> Keyword '" + keyword +
"' was found in the following command:",
force=True)
for i in out_list:
print("\t- " + i + ".")
else:
message(">> Keyword '" + keyword + "' was not found.", force=True)
def intronic(inputfile=None,
outputfile=None,
names='transcript_id',
separator="_",
intron_nb_in_name=False,
no_feature_name=False,
by_transcript=False):
"""
Extract intronic regions.
"""
message("Searching for intronic regions.")
# Need to load if the gtf comes from
# <stdin>
gtf = GTF(inputfile, check_ensembl_format=False)
if not by_transcript:
introns_bo = gtf.get_introns()
for i in introns_bo:
write_properly(chomp(str(i)), outputfile)
else:
introns_bo = gtf.get_introns(by_transcript=True,
name=names.split(","),
sep=separator,
intron_nb_in_name=intron_nb_in_name,
feat_name=not no_feature_name)
for i in introns_bo:
write_properly(chomp(str(i)), outputfile)
gc.disable()
close_properly(outputfile, inputfile)
def bed_to_gtf(inputfile=None,
outputfile=None,
ft_type="transcript",
source="Unknown"):
"""
Convert a bed file to a gtf. This will make the poor bed feel as if it was a
nice gtf (but with lots of empty fields...). May be helpful sometimes...
"""
message("Converting the bed file into GTF file.")
if inputfile.name == '<stdin>':
tmp_file = make_tmp_file(prefix="input_bed", suffix=".bed")
for i in inputfile:
write_properly(chomp(str(i)), tmp_file)
tmp_file.close()
inputfile.close()
bed_obj = BedTool(tmp_file.name)
else:
bed_obj = BedTool(inputfile.name)
n = 1
for i in bed_obj:
if i.strand == "":
i.strand = "."
if i.name == "":
i.name = str("feature_" + str(n))
if i.score == "":
i.score = "0"
if ft_type == "exon":
key_value = "gene_id \"" + i.name + "\"; " + \
"transcript_id \"" + i.name + "\"; " + \
"exon_id \"" + i.name + "\";"
elif ft_type == "gene":
key_value = "gene_id \"" + i.name + "\";"
else:
key_value = "gene_id \"" + i.name + "\"; " + \
"transcript_id \"" + i.name + "\";"
if pygtftk.utils.ADD_CHR == 1:
chrom_out = "chr" + i.chrom
else:
chrom_out = i.chrom
list_out = [
chrom_out, source, ft_type,
str(i.start + 1),
str(i.end),
str(i.score), i.strand, ".", key_value
]
write_properly("\t".join(list_out), outputfile)
n += 1
gc.disable()
close_properly(outputfile)
def find_interesting_combinations(self):
"""
MAIN FUNCTION. Will call the others.
"""
## Hardcode ignoring of Python (and by extension SKLearn) warnings
try:
previous_warning_level = os.environ["PYTHONWARNINGS"]
except:
previous_warning_level = 'default'
if utils.VERBOSITY < 2: # Only if not debugging
os.environ["PYTHONWARNINGS"] = "ignore"
#warnings.filterwarnings('ignore', module='^{}\.'.format(re.escape("sklearn")))
message("Filtering out sklearn warnings.")
## Now call the functions
self.generate_candidate_words()
self.filter_library()
self.select_best_words_from_library()
# Re-enable warnings
os.environ["PYTHONWARNINGS"] = previous_warning_level
return self.best_words
def select_best_words_from_library(self):
"""
This is step 2. Takes the library of candidates produced at step 1
and will get the best N words among it that best rebuild the original matrix.
"""
# You can't request more words than are actually present in the library,
# nor than unique elements in the data
upper_floor_words = min(self.number_of_words_in_library,
len(np.unique(self.data, axis=0)))
if self.queried_words_nb > upper_floor_words:
self.queried_words_nb = upper_floor_words
message("Requesting too many words, reducing to " +
str(self.queried_words_nb))
# NOTE It will actually be +1 to make room for the root (0,0,0,...) word, but this is added later
# Read the parameters that were supplied when creating the Modl object
best_dict = modl_subroutines.build_best_dict_from_library(
self.data,
self.library, # Data and Library of candidates
self.queried_words_nb, # N best words
self.error_function, # Potential custom error function
self.nb_threads,
self.normalize_words, # Normalize words by sum of square
self.step_2_alpha) # Sparsity control
# Final step : register the best dictionary
self.best_words = best_dict
def midpoints(inputfile=None,
outputfile=None,
ft_type="transcript",
names="transcript_id",
separator="|"):
"""
Get the midpoint coordinates for the requested feature.
"""
message("Loading input file...")
if inputfile.name == '<stdin>':
is_gtf = True
else:
region_bo = BedTool(inputfile.name)
if len(region_bo) == 0:
message("Unable to find requested regions", type="ERROR")
if region_bo.file_type == 'gff':
is_gtf = True
else:
is_gtf = False
if is_gtf:
gtf = GTF(inputfile.name, check_ensembl_format=False)
bed_obj = gtf.select_by_key("feature", ft_type).get_midpoints(
name=names.split(","), sep=separator)
for line in bed_obj:
write_properly(chomp(str(line)), outputfile)
else:
for line in region_bo:
diff = line.end - line.start
if diff % 2 != 0:
# e.g 10-13 (zero based) -> 11-13 one based
# mipoint is 12 (one-based) -> 11-12 (zero based)
# e.g 949-1100 (zero based) -> 950-1100 one based
# mipoint is 1025 (one-based) -> 1024-1025 (zero based)
# floored division (python 2)...
line.end = line.start + int(diff // 2) + 1
line.start = line.end - 1
else:
# e.g 10-14 (zero based) -> 11-14 one based
# mipoint is 12-13 (one-based) -> 11-13 (zero based)
# e.g 9-5100 (zero based) -> 10-5100 one based
# mipoint is 2555-2555 (one-based) -> 2554-2555 (zero based)
# floored division (python 2)...
# No real center. Take both
line.start = line.start + int(diff // 2) - 1
line.end = line.start + 2
outputfile.write(str(line))
gc.disable()
close_properly(outputfile, inputfile)
def negbin_pval(k, mean, var, precision=320, ft_type="Unknown"):
r"""
P-value for a negative binomial distribution of the given moments (mean, var).
This is the two-sided p-value : it will return the minimum of the left-sided
and right-sided p-value
NOTE : To prevent division by zero or negative r, if the mean is higher than
or equal to the variance, set the variance to mean + epsilon and send a warning
:param k: the critical value for which the pvalue is computed.
:param mean: The mean for the negative binomial model.
:param var: The variance for the negative binomial model.
:param precision: Floating point precision of mpmath. Should be at least 1000
:param ft_type: The name of the feature to be tested (just for meaningful messages).
>>> from pygtftk.stats.negbin_fit import negbin_pval
>>> mean = 18400
>>> var = 630200
>>> k = 65630
>>> pval = negbin_pval(k, mean, var)
>>> import math
>>> assert(math.isclose(pval,1.1999432787236828e-307))
"""
if mean < 1:
mean = 1
msg = "Computing log(p-val) for a Neg Binom with mean < 1 ; mean was set to 1 (" + ft_type + ")"
message(msg, type='WARNING')
# This is necessary, since r must be above 0.
if mean >= var:
var = mean + 1
msg = "Computing log(p-val) for a Neg Binom with mean >= var ; var was set to mean+1 (" + ft_type + ")"
message(msg, type='WARNING')
# Floating point precision of mpmath. Should be at least 320.
mpmath.mp.dps = precision
# Calculate r and p based on mean and var
r = mpmath.mpf(mean**2 / (var - mean))
p = mpmath.mpf(1 / (mean / r + 1))
# To circumvent scipy floating point precision issues, we implement a
# custom p-value calculation (see 'beta.py' for details)
mybetacalc = BetaCalculator(use_log=True,
precision=precision,
ft_type=ft_type)
incomplete_beta = mybetacalc.betainc(a=r, b=k + 1, x=p)
complete_beta = mybetacalc.beta(a=r, b=k + 1)
# Take the minimum of CDF and SF
pval = 1 - (incomplete_beta / complete_beta)
twosided_pval = min(pval, 1 - pval)
# Convert back to Python float and return
return float(twosided_pval)
def del_attr(inputfile=None,
outputfile=None,
key="transcript_id",
reg_exp=False,
invert_match=False):
"""
Delete extended attributes in the target gtf file. attr_list can be a
comma-separated list of attributes.
"""
# ----------------------------------------------------------------------
# Read the GTF and get the list of attributes
# ----------------------------------------------------------------------
gtf = GTF(inputfile, check_ensembl_format=False)
attr_list = gtf.attr_extended
# ----------------------------------------------------------------------
# If regExp, select the corresponding keys
# ----------------------------------------------------------------------
if reg_exp:
key_list = []
try:
rgxp = re.compile(key)
except:
message("Check the regular expression please.", type="ERROR")
for attr in attr_list:
if rgxp.search(attr):
key_list += [attr]
else:
key_list = key.split(",")
# ----------------------------------------------------------------------
# If invert-match select all but the selected
# ----------------------------------------------------------------------
key_to_del = []
if invert_match:
for attr in attr_list:
if attr not in key_list:
key_to_del += [attr]
else:
key_to_del = key_list
# ----------------------------------------------------------------------
# Delete the keys
# ----------------------------------------------------------------------
gtf = gtf.del_attr(feat="*", keys=",".join(key_list),
force=True).write(outputfile, gc_off=True)
close_properly(outputfile, inputfile)
def __call__(self, parser, namespace, values, option_string=None):
if not os.path.exists(CmdManager.config_dir):
message("Please run gtftk -h before adding additional plugins",
force=True)
sys.exit(0)
open(os.path.join(CmdManager.config_dir, "reload"), "w")
message("Plugins will be updated at next startup.", force=True)
sys.exit()
def dump_plugins(self):
"""Save the plugins into a pickle object."""
message("Dumping plugins", force=True)
f_handler = open(CmdManager.dumped_plugin_path, "wb")
pick = cloudpickle.CloudPickler(f_handler)
pick.dump((self.cmd_obj_list, self.parser))
f_handler.close()
def add_exon_nb(inputfile=None, outputfile=None, exon_numbering_key=None):
"""Add the exon number to each exon (based on 5' to 3' orientation)."""
message("Calling nb_exons.", type="DEBUG")
GTF(inputfile.name,
check_ensembl_format=False).add_exon_number(exon_numbering_key).write(
outputfile, gc_off=True)
close_properly(inputfile, outputfile)
def contfractbeta(self, a, b, x):
"""
Evaluates the continued fraction form of the incomplete Beta function.
Code translated from: GNU Scientific Library
Uses the modified Lentz's method.
You can see a representation of this form in the Digial Library of
Mathematical functions <https://dlmf.nist.gov/8.17#SS5.p1>.
The goal of the method is to calculate the successive 'd' terms,
separately for odd and even.
"""
a, b, x = mpmath.mpf(a), mpmath.mpf(b), mpmath.mpf(x)
num_term = 1.0
den_term = 1.0 - (a + b) * x / (a + 1.0)
den_term = 1.0 / den_term
cf = den_term
for i in range(self.itermax + 1):
k = i + 1
coeff = k * (b - k) * x / (((a - 1.0) + 2 * k) * (a + 2 * k))
# First step of the recurrence
den_term = 1.0 + coeff * den_term
num_term = 1.0 + coeff / num_term
den_term = 1.0 / den_term
delta_frac = den_term * num_term
cf *= delta_frac
coeff = -(a + k) * (a + b + k) * x / ((a + 2 * k) *
(a + 2 * k + 1.0))
# Second step
den_term = 1.0 + coeff * den_term
num_term = 1.0 + coeff / num_term
den_term = 1.0 / den_term
delta_frac = den_term * num_term
cf *= delta_frac
# Are we done ?
if (abs(delta_frac - 1.0) < 2.0 * self.epsilon):
return cf
# If failed to converge, return our best guess but send a warning
msg = 'a or b too large or given itermax too small for computing incomplete'
msg += ' beta function ; pval may be slightly erroneous for feature (' + self.ft_type + ').'
message(msg, type='WARNING')
return cf
def get_datasets(self, database=None):
message("Listing available datasets", type="DEBUG")
if database in self.databases:
self.query(query={'type': 'datasets', 'mart': database})
for i in self.response.text.split("\n"):
fields = i.split("\t")
if len(fields) > 1:
self.datasets[fields[1]] = fields[2:]
else:
message("Database not found.")
def select_by_max_exon_nb(inputfile=None, outputfile=None):
"""
Select transcripts based on the number of exons.
"""
msg = "Selecting transcript with the highest number of exon for each gene."
message(msg)
gtf = GTF(inputfile, check_ensembl_format=False).select_by_max_exon_nb()
gtf.write(outputfile, gc_off=True)
def _get_databases(self):
message("Listing available databases", type="DEBUG")
try:
self.query(query={'type': 'registry'})
except ConErr:
message("Raised a connection Error.", type="ERROR")
tree = ElementTree.fromstring(self.response.content)
for child in tree:
if child.tag == 'MartURLLocation':
self.databases += [child.attrib['name']]
def select_most_5p_tx(inputfile=None, outputfile=None, keep_gene_lines=False):
"""
Select the most 5' transcript of each gene.
"""
message("Selecting the most 5' transcript of each gene.")
gtf = GTF(inputfile)
if keep_gene_lines:
gtf = gtf.select_5p_transcript()
else:
gtf = gtf.select_5p_transcript().select_by_key("feature", "gene", 1)
gtf.write(outputfile, gc_off=True)
def _find_plugins():
message("Searching plugins", force=True)
config_file = CmdManager.config_file
# User plugins
plugin_dir_user = yaml.load(open(config_file, "r"),
Loader=yaml.FullLoader)["plugin_path"]
sys.path.append(plugin_dir_user)
plugins = sorted(os.listdir(plugin_dir_user))
plugins_user = [os.path.join(plugin_dir_user, x) for x in plugins]
# System wide plugins (those declared in the plugins directory of
# pygtftk source)
plugin_dir_base = os.path.join(pygtftk.__path__[0], "plugins")
sys.path.append(plugin_dir_base)
plugins = sorted(os.listdir(plugin_dir_base))
plugins_system = [os.path.join(plugin_dir_base, x) for x in plugins]
plugins = plugins_user + plugins_system
for plug in plugins:
if plug.endswith(".py") and plug != "__init__.py":
# Loading the plugin should force code to create
# a cmdObject that will be added to the CmdManager
# gtftk.plugins.tss_dist
module_name = re.sub("\.py$", "", plug)
module_name = re.sub("/", ".", module_name)
module_name = re.sub(".*pygtftk", "pygtftk", module_name)
try:
SourceFileLoader(module_name, plug).load_module()
except Exception as e:
message("Failed to load plugin :" + plug, type="WARNING")
print(e)
elif plug.endswith(".R"):
pass
# declare_r_cmd(plugin_path, plug)
CmdManager.reload = False
if os.path.exists(os.path.join(CmdManager.config_dir, "reload")):
os.remove(os.path.join(CmdManager.config_dir, "reload"))
def make_tmp_file_pool(prefix='tmp',
suffix='',
store=True,
dir=None):
"""
This
:Example:
>>> from pygtftk.utils import make_tmp_file_pool
>>> tmp_file = make_tmp_file_pool()
>>> assert os.path.exists(tmp_file.name)
>>> tmp_file = make_tmp_file_pool(prefix="pref")
>>> assert os.path.exists(tmp_file.name)
>>> tmp_file = make_tmp_file_pool(suffix="suf")
>>> assert os.path.exists(tmp_file.name)
"""
dir_target = None
if dir is None:
if pygtftk.utils.TMP_DIR is not None:
if not os.path.exists(pygtftk.utils.TMP_DIR):
msg = "Creating directory {d}."
message(msg.format(d=pygtftk.utils.TMP_DIR), type="INFO")
os.mkdir(pygtftk.utils.TMP_DIR)
dir_target = pygtftk.utils.TMP_DIR
else:
if not os.path.exists(dir):
msg = "Creating directory {d}."
message(msg.format(d=dir), type="INFO")
os.mkdir(dir)
dir_target = dir
tmp_file = NamedTemporaryFile(delete=False,
mode='w',
prefix=prefix + "_pygtftk_",
suffix=suffix,
dir=dir_target)
if store:
TMP_FILE_POOL_MANAGER.append(tmp_file.name)
return tmp_file
def __call__(self, minibatch_len, seed, id):
my_result = compute_all_intersections_minibatch(
self.Lr1,
self.Li1,
self.Lrs,
self.Lis,
self.all_chrom1,
self.all_chrom2,
minibatch_len,
self.use_markov_shuffling,
self.keep_intact_in_shuffling,
self.nb_threads,
seed=seed)
message("--- Minibatch nb. : " + str(id) + " is complete.")
return my_result
def select_by_tx_size(inputfile=None,
outputfile=None,
min_size=None,
max_size=None):
"""
Select features by size.
"""
msg = "Selecting 'mature/spliced transcript by size (range: [{m},{M}])."
msg = msg.format(m=str(min_size),
M=str(max_size))
message(msg)
GTF(inputfile
).select_by_transcript_size(min_size,
max_size
).write(outputfile,
gc_off=True)
def select_by_nb_exon(inputfile=None,
outputfile=None,
min_exon_number=None,
max_exon_number=None):
"""
Select transcripts based on the number of exons.
"""
msg = "Selecting transcript by exon number (range: [{m},{M}])"
msg = msg.format(m=str(min_exon_number), M=str(max_exon_number))
message(msg)
gtf = GTF(inputfile, check_ensembl_format=False).select_by_number_of_exons(
min_exon_number, max_exon_number)
gtf.write(outputfile, gc_off=True)
close_properly(outputfile, inputfile)
def exon_sizes(inputfile=None, outputfile=None, key_name=None):
"""
Add a new key to transcript features containing a comma-separated list of exon-size.
"""
gtf = GTF(inputfile)
all_tx_ids = gtf.get_tx_ids(nr=True)
tx_to_size_list = dict()
exons_starts = gtf.select_by_key("feature", "exon").extract_data(
"transcript_id,start",
as_dict_of_merged_list=True,
no_na=True,
nr=False)
if not len(exons_starts):
message("No exon found.", type="ERROR")
exons_ends = gtf.select_by_key("feature", "exon").extract_data(
"transcript_id,end", as_dict_of_merged_list=True, no_na=True, nr=False)
strands = gtf.select_by_key("feature", "transcript").extract_data(
"transcript_id,strand",
as_dict_of_values=True,
no_na=True,
nr=True,
hide_undef=True)
for tx_id in all_tx_ids:
size_list = []
for s, e in zip(exons_starts[tx_id], exons_ends[tx_id]):
size = str(int(e) - int(s) + 1)
size_list += [size]
if strands[tx_id] == "-":
size_list = reversed(size_list)
tx_to_size_list[tx_id] = ",".join(size_list)
if len(tx_to_size_list):
gtf = gtf.add_attr_from_dict(feat="transcript",
key="transcript_id",
a_dict=tx_to_size_list,
new_key=key_name)
gtf.write(outputfile, gc_off=True)
close_properly(outputfile, inputfile)
def select_by_go(inputfile=None,
outputfile=None,
go_id=None,
https_proxy=None,
http_proxy=None,
list_datasets=None,
species=None,
invert_match=False):
""" Select lines from a GTF file based using a Gene Ontology ID (e.g GO:0050789).
"""
if not go_id.startswith("GO:"):
go_id = "GO:" + go_id
is_associated = OrderedDict()
bm = Biomart(http_proxy=http_proxy,
https_proxy=https_proxy)
bm.get_datasets('ENSEMBL_MART_ENSEMBL')
if list_datasets:
for i in sorted(bm.datasets):
write_properly(i.replace("_gene_ensembl", ""), outputfile)
sys.exit()
else:
if species + "_gene_ensembl" not in bm.datasets:
message("Unknow dataset/species.", type="ERROR")
bm.query({'query': XML.format(species=species, go=go_id)})
for i in bm.response.content.decode().split("\n"):
i = i.rstrip("\n")
if i != '':
is_associated[i] = 1
gtf = GTF(inputfile)
gtf_associated = gtf.select_by_key("gene_id",
",".join(list(is_associated.keys())),
invert_match)
gtf_associated.write(outputfile,
gc_off=True)
def random_tx(inputfile=None,
outputfile=None,
max_transcript=None,
seed_value=None):
"""
Select randomly up to m transcript for each gene.
"""
message("loading the GTF.")
gtf = GTF(inputfile).select_by_key("feature", "gene", invert_match=True)
message("Getting gene_id and transcript_id")
gene2tx = gtf.extract_data("gene_id,transcript_id",
as_dict_of_merged_list=True,
no_na=True,
nr=True)
message("Selecting random transcript")
if seed_value is not None:
random.seed(seed_value, version=1)
tx_to_delete = []
for gn_id in gene2tx:
tx_list = gene2tx[gn_id]
nb_tx = len(tx_list)
max_cur = min(max_transcript, nb_tx)
pos_to_keep = random.sample(list(range(len(tx_list))), max_cur)
tx_list = [j for i, j in enumerate(tx_list) if i not in pos_to_keep]
tx_to_delete += tx_list
message("Printing results")
message("Selecting transcript.")
gtf.select_by_key("transcript_id",
",".join(tx_to_delete),
invert_match=True).write(outputfile, gc_off=True)
close_properly(outputfile, inputfile)
def join_multi_file(inputfile=None,
outputfile=None,
target_feature=None,
key_to_join=None,
matrix_files=()):
"""
Join attributes from a set of tabulated files.
"""
# -----------------------------------------------------------
# load the GTF
# -----------------------------------------------------------
gtf = GTF(inputfile, check_ensembl_format=False)
# -----------------------------------------------------------
# Check target feature
# -----------------------------------------------------------
feat_list = gtf.get_feature_list(nr=True)
if target_feature is not None:
target_feature_list = target_feature.split(",")
for i in target_feature_list:
if i not in feat_list + ["*"]:
message("Feature " + i + " not found.", type="ERROR")
else:
target_feature = ",".join(feat_list)
# -----------------------------------------------------------
# Do it
# -----------------------------------------------------------
for join_file in matrix_files:
gtf = gtf.add_attr_from_matrix_file(feat=target_feature,
key=key_to_join,
inputfile=join_file.name)
gtf.write(outputfile, gc_off=True)
gc.disable()
close_properly(outputfile, inputfile)
def intergenic(inputfile=None, outputfile=None, chrom_info=None):
"""
Extract intergenic regions.
"""
message("Searching for intergenic regions.")
gtf = GTF(inputfile)
intergenic_regions = gtf.get_intergenic(chrom_info)
nb_intergenic_region = 1
for i in intergenic_regions:
i.name = "region_" + str(nb_intergenic_region)
write_properly(chomp(str(i)), outputfile)
nb_intergenic_region += 1
gc.disable()
close_properly(outputfile, inputfile)
def parse_cmd_args(cls):
""" Parse arguments of all declared commands."""
CmdManager.args = cls.parser.parse_args(None)
args = CmdManager.args
cmd_name = args.command
if cmd_name is None:
message("Please provide a subcommand or argument (e.g. -h)",
type="WARNING",
force=True)
CmdManager.parser.print_help()
exit(0)
lang = cls.cmd_obj_list[cmd_name].lang
if lang == 'Python':
if args.tmp_dir is not None:
if not os.path.exists(args.tmp_dir):
msg = "Creating directory {d}."
message(msg.format(d=args.tmp_dir), type="INFO")
mkdir_p(args.tmp_dir)
if not os.path.isdir(args.tmp_dir):
msg = "{d} is not a directory."
message(msg.format(d=args.tmp_dir), type="ERROR")
pygtftk.utils.TMP_DIR = args.tmp_dir
return args
def query(self, query):
message("Sending query", type="DEBUG")
self.response = requests.get(self.url, query, proxies=self.proxies)
message("Checking http response", type="DEBUG")
if self.response.status_code != requests.codes.ok:
msg = "HTTP response status code: {c}. {m}"
msg = msg.format(c=str(self.response.status_code),
m=self.response.reason)
message(msg, type="ERROR")
msg = "([ \.\w]+ service you requested is currently unavailable[ \.\w]+)"
hit = re.search(msg, self.response.text)
if hit:
msg = re.search(msg, self.response.text).group(1)
message(msg.lstrip().rstrip(), type="WARNING")
message(
"More information about this downtime "
"may be available on http://www.ensembl.info/",
type="ERROR")
def nb_exons(inputfile=None,
outputfile=None,
key_name=None,
text_format=False):
"""
Count the number of exons in the gtf file.
"""
gtf = GTF(inputfile)
n_exons = defaultdict(int)
# -------------------------------------------------------------------------
# Computing number of exon for each transcript in input GTF file
#
# -------------------------------------------------------------------------
message("Computing number of exons for each transcript in input GTF file.")
exon = gtf.select_by_key("feature", "exon")
fields = exon.extract_data("transcript_id")
for i in fields:
tx_id = i[0]
n_exons[tx_id] += 1
if text_format:
for tx_id in n_exons:
outputfile.write(tx_id + "\t" + str(n_exons[tx_id]) +
"\ttranscript\n")
else:
if len(n_exons):
gtf = gtf.add_attr_from_dict(feat="transcript",
key="transcript_id",
a_dict=n_exons,
new_key=key_name)
gtf.write(outputfile, gc_off=True)
close_properly(outputfile, inputfile)
def produce_dot_for_node(node, graph):
for c in node.children:
node_name = node_to_combi_string(node, features_names)
child_name = node_to_combi_string(c, features_names)
message("Drawing " + node_name + ' --> ' + child_name, type='DEBUG')
## Add nodes
# Only add node if not already present of course.
# If present, the graph's 'body' contains the combi string prefixed with a tab character
## Parent
if not ('\t' + node_name in s.body):
# print(combi_string, node.s, node.pval, node.fc)
graph.node(node_name, format_node_string(node_name, node.s, node.pval, node.fc))
# Child
if not ('\t' + child_name in s.body):
graph.node(child_name, format_node_string(child_name, c.s, c.pval, c.fc))
graph.edge(node_name + ':s', child_name + ':n')
return 1