Sequence Words With Regex

I search the sequence: nunca[ADV+NEG+CIRC] más[ADV+comp+CIRC] compraré[V+H_PREDICAT_ACTION] and nunca más compraré My script: corpus = 'Me[Unknown] temo[Unknown] que[Unkno

Solution 1:

If the searched substrings are in arbitrary order - use the following: re.findall() approach:

corpus = "Me[Unknown] temo[Unknown] que[Unknown] buscare[Unknown] \
otras[Unknown] opciones[Unknown] esta[Unknown] nunca[ADV+NEG+CIRC] \
más[ADV+comp+PADV+H_CIRCONSTANT_QUANTITE] compraré[V+H_PREDICAT_ACTION]"

result = ' '.join(i[0] for i in re.findall(r'(\w+)\[[^][]*(AD|V)\+[^][]*\]', corpus, re.M | re.UNICODE))
print(result)

The output:

nunca más compraré

---

regex pattern explanation:

• (\w+) - match a word(alphanumeric sequence) (for ex. nunca). Placed into the first captured group (...)

• \[ - match opening square bracket [ literally

• [^][]* - match one or many characters except square brackets ][

• (AD|V) - alternation group, match either AD or V key

• \] - match closing square bracket ] literally

for ex. \[[^][]*(AD|V)\+[^][]*\] will match [ADV+NEG+CIRC]

----------

If the order of sequences is strict - use re.sub() function instead re.findall() to remove all parenthetical sequences:

corpus = "Me[Unknown] temo[Unknown] que[Unknown] buscare[Unknown] \
otras[Unknown] opciones[Unknown] esta[Unknown] nunca[ADV+NEG+CIRC] \
más[ADV+comp+PADV+H_CIRCONSTANT_QUANTITE] compraré[V+H_PREDICAT_ACTION]"

result = re.sub(r'\[[^][]+\]', '', corpus, re.M | re.UNICODE)
print(result)

The output:

Me temo que buscare otras opciones esta nunca más compraré

---

To extract the last 3 words:

print(' '.join(result.split()[-3:]))    # nunca más compraré