def proctree(r, i=1):
tg.map_stree(g, r)
taxids = set()
for lf in r.leaves():
taxids.update(lf.taxid_rootpath)
taxg = tg.taxid_new_subgraph(g, taxids)
verts = taxg.new_vertex_property('bool')
edges = taxg.new_edge_property('bool')
# add stree's nodes and branches into taxonomy graph
tg.merge_stree(taxg, r, i, verts, edges)
# next, add taxonomy edges to taxg connecting 'incertae sedis'
# leaves in stree to their containing taxa
for lf in r.leaves():
if lf.taxid and lf.taxid in taxg.taxid_vertex and lf.incertae_sedis:
taxv = taxg.taxid_vertex[lf.taxid]
ev = taxg.edge(taxv, lf.v, True)
if ev:
assert len(ev)==1
e = ev[0]
else:
e = taxg.add_edge(taxv, lf.v)
taxg.edge_in_taxonomy[e] = 1
# make a view of taxg that keeps only the vertices and edges traced by
# the source tree
gv = tg.graph_view(taxg, vfilt=verts, efilt=edges)
gv.vertex_strees = taxg.vertex_strees
gv.edge_strees = taxg.edge_strees
# the following code sets up the visualization
ecolor = taxg.new_edge_property('string')
for e in taxg.edges():
est = taxg.edge_strees[e]
eit = taxg.edge_in_taxonomy[e]
if len(est) and not eit: ecolor[e] = 'blue'
elif len(est) and eit: ecolor[e] = 'green'
else: ecolor[e] = 'yellow'
ewidth = taxg.new_edge_property('int')
for e in taxg.edges():
est = taxg.edge_strees[e]
if len(est): ewidth[e] = 3
else: ewidth[e] = 1
vcolor = taxg.new_vertex_property('string')
for v in taxg.vertices():
if not taxg.vertex_in_taxonomy[v]: vcolor[v] = 'blue'
else: vcolor[v] = 'green'
vsize = taxg.new_vertex_property('int')
for v in taxg.vertices():
if taxg.vertex_in_taxonomy[v] or v.out_degree()==0:
vsize[v] = 4
else: vsize[v] = 2
pos, pin = tg.layout(taxg, gv, gv.root, sfdp=True, deg0=195.0,
degspan=150.0, radius=400)
for v in gv.vertices(): pin[v] = 1
for e in taxg.edges():
src = e.source()
tgt = e.target()
if not verts[src]:
verts[src] = 1
pos[src] = [0.0, 0.0]
vcolor[src] = 'red'
if not verts[tgt]:
verts[tgt] = 1
pos[tgt] = [0.0, 0.0]
vcolor[tgt] = 'red'
if not edges[e]:
edges[e] = 1
ecolor[e] = 'red'
ewidth[e] = 1.0
gv.wt[e] = 1.0
pos = gt.sfdp_layout(gv, pos=pos, pin=pin, eweight=gv.wt, multilevel=False)
nodes = []
links = []
idx = {}
xmin = min([ pos[x][0] for x in gv.vertices() ])
ymin = min([ pos[x][1] for x in gv.vertices() ])
for x in gv.vertices(): pos[x] = [pos[x][0]-xmin, pos[x][1]-ymin]
for i,v in enumerate(gv.vertices()):
idx[int(v)] = i
taxid = gv.vertex_taxid[v]
try:
name = gv.vertex_name[v]
except:
name = gv.taxid_name(taxid) if taxid else ''#'node%s' % int(v)
isleaf = v.out_degree()==0
d = dict(label=name, isleaf=isleaf, strees=list(gv.vertex_strees[v]),
altlabel=name)
if taxid: d['taxid'] = taxid
## if dist: d['dist'] = dist[v]
if pos and pos[v]:
x, y = pos[v]
d['x'] = x; d['y'] = y
d['fixed'] = True
d['color'] = vcolor[v]
d['size'] = vsize[v]
nodes.append(d)
for e in gv.edges():
source = idx[int(e.source())]
target = idx[int(e.target())]
strees = gv.edge_strees[e]
d = dict(source=source, target=target, strees = list(strees),
taxedge=bool(gv.edge_in_taxonomy[e]))
d['color'] = ecolor[e]
d['width'] = ewidth[e]
links.append(d)
return dict(nodes=nodes, links=links)
def proc(g, line, merged, probfile, outfile):
pbtree, s = line.split()
print 'processing', pbtree
r = ivy.newick.parse(s) # the root node of the tree of interest
lvs = r.leaves()
rps = [] # rootpaths of leaf nodes, where each rootpath is a list
# of taxids from leaf to root
leaf_tid_counts = tg.Counter()
try:
for lf in lvs:
# assign/compute attributes of leaves
w = lf.label.split('_')
lf.gi = int(w[-2][2:])
lf.taxid = int(w[-1][2:])
leaf_tid_counts[lf.taxid] += 1
if lf.taxid not in g.taxid_vertex and lf.taxid in merged:
lf.taxid = merged[lf.taxid]
## lf.taxv = g.taxid_vertex[lf.taxid]
taxv = g.taxid_vertex[lf.taxid]
lf.taxid_next, lf.taxid_back = g.hindex[taxv]
lf.taxid_rootpath = tg.taxid_rootpath(g, lf.taxid)
for i, x in enumerate(lf.taxid_rootpath):
if x not in g.taxid_vertex and x in merged:
lf.taxid_rootpath[i] = merged[x]
rps.append(lf.taxid_rootpath)
except:
print '!!! problem assigning leaf taxids'
probfile.write('%s\n' % pbtree)
#return []
r.mrca = tg.rootpath_mrca(rps) # taxid of mrca of all tree's leaves
taxids = set()
for rp in rps:
# trim rootpaths: make them terminate with mrca
while 1:
if rp[-1] == r.mrca: break
else: rp.pop()
assert rp
taxids.update(rp)
# create a taxonomy (sub)graph of only those taxids represented in r
## taxidsubg = tg.taxid_subgraph(g, taxids)
taxidsubg = tg.taxid_new_subgraph(g, taxids)
taxidsubg.vfilt = taxidsubg.new_vertex_property('bool')
## r.taxv = taxidsubg.taxid_vertex[r.mrca]
# no need to check for convexity for singleton tip taxa
for x in [ taxidsubg.taxid_vertex[lf.taxid] for lf in lvs
if leaf_tid_counts[lf.taxid]==1 ]:
taxidsubg.vfilt[x] = 0
# an undirected graph having the same topology as r, used for
# checking convexity of taxa
treegraph = tg.gt.Graph(directed=False)
treegraph.mrca = r.mrca
print 'mrca:', g.taxid_name(r.mrca)
treegraph.vertex_taxid = tg.get_or_create_vp(treegraph, 'taxid', 'int')
## treegraph.vertex_taxv = tg.get_or_create_vp(treegraph, 'taxv', 'int')
v2lf = {}
N = len(r)
verts = treegraph.add_vertex(N)
for n in r: # for each node in r
# store its treegraph vertex
n.v = verts.next()
if not n.children:
treegraph.vertex_taxid[n.v] = n.taxid
## treegraph.vertex_taxv[n.v] = int(n.taxv)
v2lf[n.v] = n
if n.parent:
treegraph.add_edge(n.parent.v, n.v)
treegraph_leaves = [ x for x in treegraph.vertices() if x.out_degree()==1 ]
convex = {} # for storing the convex subgraphs
def traverse(taxv):
"""
`taxv` is a vertex in the taxonomy graph. This function checks whether
it is convex in `treegraph`; if yes, stores the info in
`convex`; if no, it recursively checks descendants of `taxv` for
convexity
"""
tid = taxidsubg.vertex_taxid[taxv]
print 'checking', tid, taxidsubg.vertex_name[taxv]
p, c = color_vertices(g, treegraph, tid)
if len(c)==1 and len(c[1])==1: # taxv/tid is convex
print '...success'
rv = c[1][0] # rv is the root of the convex subtree
treegraph.set_vertex_filter(p)
## lvs = [ x for x in treegraph.vertices() if x.out_degree()==1 ]
lvs = [ x for x in treegraph_leaves if p[x] ]
if len(lvs) > 2:
# we are only interested in convex subgraphs having
# more than 2 leaves
rootpaths = []
for lf in lvs:
ti = treegraph.vertex_taxid[lf]
tv = taxidsubg.taxid_vertex[ti]
if not taxidsubg.incertae_sedis[tv]:
rootpaths.append(tg.taxid_rootpath(taxidsubg, ti))
if rootpaths:
mrca = tg.rootpath_mrca(rootpaths)
print 'traverse: mrca', mrca
ancv = [taxidsubg.taxid_vertex[mrca]]
while ancv[-1] != taxv:
# STRANGE EDGE CASES HERE
try: ancv.append(ancv[-1].in_neighbours().next())
except StopIteration: pass
k = '.'.join([ str(taxidsubg.vertex_taxid[x])
for x in ancv ])
convex[k] = (rv, p)
treegraph.set_vertex_filter(None)
else:
treegraph.set_vertex_filter(None)
for n in taxv.out_neighbours():
traverse(n)
for v in taxidsubg.root.out_neighbours(): traverse(v)
## print 'done'
def make_newick(root, seen):
children = [ x for x in root.out_neighbours() if x not in seen ]
if children:
seen.update(children)
s = '(%s)' % ','.join(
[ make_newick(c, seen) for c in children ]
)
else:
s = v2lf[root].label.replace(',','').replace('(','').replace(')','')
return s
newicks = []
for k, (root, p) in convex.items():
treegraph.set_vertex_filter(p)
s = make_newick(root, set([root]))
treegraph.set_vertex_filter(None)
names = ','.join([ g.taxid_name(int(x)) for x in k.split('.') ])
outfile.write('%s\t%s\t%s\t%s;\n' % (pbtree, k, names, s))
print 'wrote subtree:', names
for n in r.postiter():
n.parent = None; del n.children
ivy.tree.index(r)
for n in r:
if n.isleaf:
v = n.label.split('_')
n.snode_id = int(v[0])
n.taxid = int(v[1]) if (len(v) > 1 and v[1]
and v[1] != 'None') else None
else:
n.snode_id = int(n.label)
r.stree = stree
tg.map_stree(g, r)
taxids = set()
for lf in r.leaves():
taxids.update(lf.taxid_rootpath)
taxg = tg.taxid_new_subgraph(g, taxids)
# taxg is a new graph containing only the taxids in stree
# these properties will store the vertices and edges that are traced
# by r
verts = taxg.new_vertex_property('bool')
edges = taxg.new_edge_property('bool')
# add stree's nodes and branches into taxonomy graph
tg.merge_stree(taxg, r, stree, verts, edges)
# verts and edges now filter the paths traced by r in taxg
# next, add taxonomy edges to taxg connecting 'incertae sedis'
# leaves in stree to their containing taxa
for lf in r.leaves():
if lf.taxid and lf.taxid in taxg.taxid_vertex and lf.incertae_sedis:
ivy.tree.index(r)
for n in r:
if n.isleaf:
v = n.label.split('_')
n.snode_id = int(v[0])
n.taxid = int(v[1]) if (len(v)>1 and
v[1] and v[1] != 'None') else None
else:
n.snode_id = int(n.label)
r.stree = stree
tg.map_stree(g, r)
taxids = set()
for lf in r.leaves():
taxids.update(lf.taxid_rootpath)
taxg = tg.taxid_new_subgraph(g, taxids)
# taxg is a new graph containing only the taxids in stree
# these properties will store the vertices and edges that are traced
# by r
verts = taxg.new_vertex_property('bool')
edges = taxg.new_edge_property('bool')
# add stree's nodes and branches into taxonomy graph
tg.merge_stree(taxg, r, stree, verts, edges)
# verts and edges now filter the paths traced by r in taxg
# next, add taxonomy edges to taxg connecting 'incertae sedis'
# leaves in stree to their containing taxa
for lf in r.leaves():
if lf.taxid and lf.taxid in taxg.taxid_vertex and lf.incertae_sedis:
def build_json(choice):
if choice == "1":
## Loads a graph with the OTT taxonomy
taxonomy="ott"
print "Loading OTT taxonomy into graph..."
g = tg.load_taxonomy_graph('taxonomy/ott2.2/ott2.2.xml.gz')
print "OTT taxonomy Graph loaded successfully."
print "Loading ott-treecache file..."
datafile = open('trees/ott-treecache.txt', 'r') #read in the treecache file
print "Loaded."
elif choice == "2":
taxonomy="ncbi"
print "Loading NCBI taxonomy into graph..."
g = tg.load_taxonomy_graph('taxonomy/ncbi/ncbi.xml.gz')
print "NCBI taxonomy Graph loaded successfully."
print "Loading ncbi-treecache file..."
datafile = open('trees/ncbi-treecache.txt', 'r') #read in the treecache file
print "Loaded."
data = []
errors = []
blacklist = []
## Loop all of the entries in the treecache.txt file and assign them to data.
for row in datafile:
data.append(row)
#print row
## Creates a Tree Blacklist that will ignore problematic trees that cause crashes based on strange formatting issues until then can be resolved.
print "Loading tree blacklist..."
tree_blacklist = open('trees/tree_blacklist.txt', 'r') #read in the tree blacklist file
print "Loaded."
## Loop all of the entries in the tree_blacklist.txt file and assign them to blacklist.
for tree in tree_blacklist:
blacklist.append(tree.strip())
rowcount = 0
for row in data: #iterate through each unique stree id in the file allowing the code below to generate the graph, write the JSON and save the file
active_tree = row.split(":") #split the row from treecache into tree id and newick string tree
if active_tree[0] in blacklist: ## if a tree is in the blacklist, ignore it.
print ("Tree %s is being ignored as it is black listed." % active_tree[0])
else:
stree = int(active_tree[0]) # convert tree id string into int
r = ivy.tree.read(active_tree[1].replace("?", "")) #read the tree, also replacing an extraneous ? characters
leafcount = 0
r.ladderize()
ivy.tree.index(r)
for n in r:
if n.isleaf:
leafcount = leafcount + 1
v = n.label.split('_')
n.snode_id = int(v[0])
n.taxid = int(v[1]) if (len(v)>1 and
v[1] and v[1] != 'None') else None
else:
n.snode_id = int(n.label)
if leafcount <= 5000: #check to prune trees that have more than 5000 leaves. They will not display correctly in graph form.
try: #used to catch all errors from incorrectly formatted trees (ie: ? characters, and other issues)
r.stree = stree
### ADD CODE HERE TO SKIP TREES WITH MORE THAN 5000 leaves
tg.map_stree(g, r)
taxids = set()
for lf in r.leaves():
taxids.update(lf.taxid_rootpath)
taxg = tg.taxid_new_subgraph(g, taxids)
# taxg is a new graph containing only the taxids in stree
# these properties will store the vertices and edges that are traced
# by r
verts = taxg.new_vertex_property('bool')
edges = taxg.new_edge_property('bool')
# add stree's nodes and branches into taxonomy graph
tg.merge_stree(taxg, r, stree, verts, edges)
# verts and edges now filter the paths traced by r in taxg
# next, add taxonomy edges to taxg connecting 'incertae sedis'
# leaves in stree to their containing taxa
for lf in r.leaves():
if lf.taxid and lf.incertae_sedis:
taxv = taxg.taxid_vertex[lf.taxid]
ev = taxg.edge(taxv, lf.v, True)
if ev:
assert len(ev)==1
e = ev[0]
else:
e = taxg.add_edge(taxv, lf.v)
taxg.edge_in_taxonomy[e] = 1
# make a view of taxg that keeps only the vertices and edges traced by
# the source tree
gv = tg.graph_view(taxg, vfilt=verts, efilt=edges)
gv.vertex_strees = taxg.vertex_strees
gv.edge_strees = taxg.edge_strees
# the following code sets up the visualization
ecolor = taxg.new_edge_property('string')
for e in taxg.edges():
est = taxg.edge_strees[e]
eit = taxg.edge_in_taxonomy[e]
if len(est) and not eit: ecolor[e] = 'blue'
elif len(est) and eit: ecolor[e] = 'green'
else: ecolor[e] = 'yellow'
ewidth = taxg.new_edge_property('int')
for e in taxg.edges():
est = taxg.edge_strees[e]
if len(est): ewidth[e] = 3
else: ewidth[e] = 1
vcolor = taxg.new_vertex_property('string')
for v in taxg.vertices():
if not taxg.vertex_in_taxonomy[v]: vcolor[v] = 'blue'
else: vcolor[v] = 'green'
vsize = taxg.new_vertex_property('int')
for v in taxg.vertices():
if taxg.vertex_in_taxonomy[v] or v.out_degree()==0:
vsize[v] = 8
else: vsize[v] = 2
pos, pin = tg.layout(taxg, gv, gv.root, sfdp=True, deg0=195.0,
degspan=150.0, radius=400)
for v in gv.vertices(): pin[v] = 1
for e in taxg.edges():
src = e.source()
tgt = e.target()
if not verts[src]:
verts[src] = 1
pos[src] = [0.0, 0.0]
vcolor[src] = 'red'
if not verts[tgt]:
verts[tgt] = 1
pos[tgt] = [0.0, 0.0]
vcolor[tgt] = 'red'
if not edges[e]:
edges[e] = 1
ecolor[e] = 'red'
ewidth[e] = 1.0
gv.wt[e] = 1.0
pos = gt.sfdp_layout(gv, pos=pos, pin=pin, eweight=gv.wt, multilevel=False)
### Use function in TreeGraph.py to parse Graph(gv) into JSON
print "Generating JSON..."
result = tg.graph_json(gv, pos=pos, ecolor=ecolor, ewidth=ewidth, vcolor=vcolor, vsize=vsize)
result = result[1:] #strip the original { from the json so we can insert the time stamp
date = time.strftime("%Y%m%d%I%M%S") # grab the system date for the filename and convert it to a string
treeid = str(stree) # convert stree int into a string
timestamp = "{\"timestamp\": \"%s\", " %date
final_result = timestamp+result # add date to first line of json file for later parsing
path = str(os.path.dirname(os.path.realpath(__file__)))
path = path[:-8]
path = "%s//%s/" % (path, taxonomy) # build the full path to write the file too
filename = "%stree_%s.JSON" % (path, treeid) # build the full file_name for writing
if not os.path.exists(path): ## if directory doesn't exist, create it.
os.makedirs(path)
f = open(filename, 'w')
f.write(final_result)
f.close
print "Done."
rowcount = rowcount + 1
except: # catch *all* exceptions
e = sys.exc_info()[0]
e = str(e)
treeid = str(stree)
print ("Error: %s</p>" % e)
errorstring = "Error: " + e + " on Tree: " + treeid # rough hack to store trees with errors and the general error
errors.append(errorstring) # store all of the error strings
rowcount = rowcount + 1
continue ## continue converting the rest of the trees into JSON even if a specific tree has errors
else:
print "Tree has more than 5000 leaves. No graph will be generated."
print "JSON Generation Complete."
## write the error strings to a log file for review later
if errors:
with open("error_log.txt", "w+") as error_log:
pickle.dump(errors, error_log)
def build_json(choice):
if choice == "1":
## Loads a graph with the OTT taxonomy
taxonomy = "ott"
print "Loading OTT taxonomy into graph..."
g = tg.load_taxonomy_graph('taxonomy/ott2.2/ott2.2.xml.gz')
print "OTT taxonomy Graph loaded successfully."
print "Loading ott-treecache file..."
datafile = open('trees/ott-treecache.txt',
'r') #read in the treecache file
print "Loaded."
elif choice == "2":
taxonomy = "ncbi"
print "Loading NCBI taxonomy into graph..."
g = tg.load_taxonomy_graph('taxonomy/ncbi/ncbi.xml.gz')
print "NCBI taxonomy Graph loaded successfully."
print "Loading ncbi-treecache file..."
datafile = open('trees/ncbi-treecache.txt',
'r') #read in the treecache file
print "Loaded."
data = []
errors = []
blacklist = []
## Loop all of the entries in the treecache.txt file and assign them to data.
for row in datafile:
data.append(row)
#print row
## Creates a Tree Blacklist that will ignore problematic trees that cause crashes based on strange formatting issues until then can be resolved.
print "Loading tree blacklist..."
tree_blacklist = open('trees/tree_blacklist.txt',
'r') #read in the tree blacklist file
print "Loaded."
## Loop all of the entries in the tree_blacklist.txt file and assign them to blacklist.
for tree in tree_blacklist:
blacklist.append(tree.strip())
rowcount = 0
for row in data: #iterate through each unique stree id in the file allowing the code below to generate the graph, write the JSON and save the file
active_tree = row.split(
":"
) #split the row from treecache into tree id and newick string tree
if active_tree[
0] in blacklist: ## if a tree is in the blacklist, ignore it.
print("Tree %s is being ignored as it is black listed." %
active_tree[0])
else:
stree = int(active_tree[0]) # convert tree id string into int
r = ivy.tree.read(active_tree[1].replace(
"?",
"")) #read the tree, also replacing an extraneous ? characters
leafcount = 0
r.ladderize()
ivy.tree.index(r)
for n in r:
if n.isleaf:
leafcount = leafcount + 1
v = n.label.split('_')
n.snode_id = int(v[0])
n.taxid = int(v[1]) if (len(v) > 1 and v[1]
and v[1] != 'None') else None
else:
n.snode_id = int(n.label)
if leafcount <= 5000: #check to prune trees that have more than 5000 leaves. They will not display correctly in graph form.
try: #used to catch all errors from incorrectly formatted trees (ie: ? characters, and other issues)
r.stree = stree
### ADD CODE HERE TO SKIP TREES WITH MORE THAN 5000 leaves
tg.map_stree(g, r)
taxids = set()
for lf in r.leaves():
taxids.update(lf.taxid_rootpath)
taxg = tg.taxid_new_subgraph(g, taxids)
# taxg is a new graph containing only the taxids in stree
# these properties will store the vertices and edges that are traced
# by r
verts = taxg.new_vertex_property('bool')
edges = taxg.new_edge_property('bool')
# add stree's nodes and branches into taxonomy graph
tg.merge_stree(taxg, r, stree, verts, edges)
# verts and edges now filter the paths traced by r in taxg
# next, add taxonomy edges to taxg connecting 'incertae sedis'
# leaves in stree to their containing taxa
for lf in r.leaves():
if lf.taxid and lf.incertae_sedis:
taxv = taxg.taxid_vertex[lf.taxid]
ev = taxg.edge(taxv, lf.v, True)
if ev:
assert len(ev) == 1
e = ev[0]
else:
e = taxg.add_edge(taxv, lf.v)
taxg.edge_in_taxonomy[e] = 1
# make a view of taxg that keeps only the vertices and edges traced by
# the source tree
gv = tg.graph_view(taxg, vfilt=verts, efilt=edges)
gv.vertex_strees = taxg.vertex_strees
gv.edge_strees = taxg.edge_strees
# the following code sets up the visualization
ecolor = taxg.new_edge_property('string')
for e in taxg.edges():
est = taxg.edge_strees[e]
eit = taxg.edge_in_taxonomy[e]
if len(est) and not eit: ecolor[e] = 'blue'
elif len(est) and eit: ecolor[e] = 'green'
else: ecolor[e] = 'yellow'
ewidth = taxg.new_edge_property('int')
for e in taxg.edges():
est = taxg.edge_strees[e]
if len(est): ewidth[e] = 3
else: ewidth[e] = 1
vcolor = taxg.new_vertex_property('string')
for v in taxg.vertices():
if not taxg.vertex_in_taxonomy[v]: vcolor[v] = 'blue'
else: vcolor[v] = 'green'
vsize = taxg.new_vertex_property('int')
for v in taxg.vertices():
if taxg.vertex_in_taxonomy[v] or v.out_degree() == 0:
vsize[v] = 8
else:
vsize[v] = 2
pos, pin = tg.layout(taxg,
gv,
gv.root,
sfdp=True,
deg0=195.0,
degspan=150.0,
radius=400)
for v in gv.vertices():
pin[v] = 1
for e in taxg.edges():
src = e.source()
tgt = e.target()
if not verts[src]:
verts[src] = 1
pos[src] = [0.0, 0.0]
vcolor[src] = 'red'
if not verts[tgt]:
verts[tgt] = 1
pos[tgt] = [0.0, 0.0]
vcolor[tgt] = 'red'
if not edges[e]:
edges[e] = 1
ecolor[e] = 'red'
ewidth[e] = 1.0
gv.wt[e] = 1.0
pos = gt.sfdp_layout(gv,
pos=pos,
pin=pin,
eweight=gv.wt,
multilevel=False)
### Use function in TreeGraph.py to parse Graph(gv) into JSON
print "Generating JSON..."
result = tg.graph_json(gv,
pos=pos,
ecolor=ecolor,
ewidth=ewidth,
vcolor=vcolor,
vsize=vsize)
result = result[
1:] #strip the original { from the json so we can insert the time stamp
date = time.strftime(
"%Y%m%d%I%M%S"
) # grab the system date for the filename and convert it to a string
treeid = str(stree) # convert stree int into a string
timestamp = "{\"timestamp\": \"%s\", " % date
final_result = timestamp + result # add date to first line of json file for later parsing
path = str(os.path.dirname(os.path.realpath(__file__)))
path = path[:-8]
path = "%s//%s/" % (
path, taxonomy
) # build the full path to write the file too
filename = "%stree_%s.JSON" % (
path, treeid) # build the full file_name for writing
if not os.path.exists(
path): ## if directory doesn't exist, create it.
os.makedirs(path)
f = open(filename, 'w')
f.write(final_result)
f.close
print "Done."
rowcount = rowcount + 1
except: # catch *all* exceptions
e = sys.exc_info()[0]
e = str(e)
treeid = str(stree)
print("Error: %s</p>" % e)
errorstring = "Error: " + e + " on Tree: " + treeid # rough hack to store trees with errors and the general error
errors.append(
errorstring) # store all of the error strings
rowcount = rowcount + 1
continue ## continue converting the rest of the trees into JSON even if a specific tree has errors
else:
print "Tree has more than 5000 leaves. No graph will be generated."
print "JSON Generation Complete."
## write the error strings to a log file for review later
if errors:
with open("error_log.txt", "w+") as error_log:
pickle.dump(errors, error_log)
